diff --git a/plugins/moody-s/skills/moody-s-earnings-brief/SKILL.md b/plugins/moody-s/skills/moody-s-earnings-brief/SKILL.md
index 3de3dda5..fb92d92e 100644
--- a/plugins/moody-s/skills/moody-s-earnings-brief/SKILL.md
+++ b/plugins/moody-s/skills/moody-s-earnings-brief/SKILL.md
@@ -56,20 +56,17 @@ This skill may **only** extract information from the following two sources:
    is older than **100 days** from today, discard it and treat the transcript as unavailable
    for that company. Do not fall back to any other source — handle the company per the
    missing-transcript policy (note in the Executive Summary, omit from every other section).
-   These rules apply to **every** section of the report. No section may be enriched with credit,
-   sector, ESG, peer, filings, or news data.
+These rules apply to **every** section of the report. No section may be enriched with credit,
+sector, ESG, peer, filings, or news data.
 
 ## Bundled files
-
 - `assets/template.html` — self-contained static report shell (CSS + named section placeholders).
   No embedded data, no inline script. Treat this file as the **read-only structural reference**:
   read it, fill it in, and write the complete filled document to a new `.html` file in
   `/mnt/user-data/outputs/`.
-
 ## Template (shared)
 
 Before emitting the HTML report, **read both**:
-
 1. [`skills/shared/template/SKILL.md`](../shared/template/SKILL.md) — authoring rules (which
    classes / snippets are owned by the shared layer, allowed per-skill overrides, outlook-badge
    usage).
@@ -77,12 +74,12 @@ Before emitting the HTML report, **read both**:
    canonical CSS (inside `<style id="shared-template-css">`) and literal HTML markup snippets
    (inside `<template>` tags) for the document head, cover, TOC, section block, sources-section
    wrapper, footer, and outlook-badge.
-   **Lookup order — always check the shared template before inventing.** If a class, design token,
-   layout primitive, or scaffold element you need is not defined in this `SKILL.md` or already
-   present in this skill's `assets/template.html`, the shared template skill is authoritative. Do
-   not invent CSS, HTML scaffolds, or design tokens that the shared skill already provides; do not
-   silently restyle anything the shared skill owns (cover, TOC, section, sources-section wrapper,
-   footer, outlook-badge, design tokens, reset, body / page base).
+**Lookup order — always check the shared template before inventing.** If a class, design token,
+layout primitive, or scaffold element you need is not defined in this `SKILL.md` or already
+present in this skill's `assets/template.html`, the shared template skill is authoritative. Do
+not invent CSS, HTML scaffolds, or design tokens that the shared skill already provides; do not
+silently restyle anything the shared skill owns (cover, TOC, section, sources-section wrapper,
+footer, outlook-badge, design tokens, reset, body / page base).
 
 At emit time, copy the **contents** (not the `<style>` wrapper) of `<style id="shared-template-css">`
 from the shared asset into the parent template's reserved marker region between the CSS-comment
@@ -103,19 +100,18 @@ solid-fill or inline-color overrides.
 
 Before emitting any `[n]` reference inline, any per-section recap block, or the end-of-document
 Citations block, **read both**:
-
 1. [`skills/shared/citations/SKILL.md`](../shared/citations/SKILL.md) — authoring rules
    (numbering, hyperlinking, source data shape, carve-outs).
 2. [`skills/shared/citations/assets/template.html`](../shared/citations/assets/template.html) —
    canonical CSS (inside `<style id="shared-citations-css">`) and literal HTML markup snippets
    (inside `<template>` tags) for inline references, the end-of-document Citations block, and
    the optional `.section-citations` recap.
-   At emit time, copy the **contents** (not the wrapper) of `<style id="shared-citations-css">`
-   from the shared asset into the parent template's reserved marker region, located inside
-   `assets/template.html` between the CSS-comment markers
-   `/* BEGIN shared-citations-css … */` and `/* END shared-citations-css */`. The parent
-   template no longer carries duplicated citation CSS — those rules ship only in the shared
-   asset.
+At emit time, copy the **contents** (not the wrapper) of `<style id="shared-citations-css">`
+from the shared asset into the parent template's reserved marker region, located inside
+`assets/template.html` between the CSS-comment markers
+`/* BEGIN shared-citations-css … */` and `/* END shared-citations-css */`. The parent
+template no longer carries duplicated citation CSS — those rules ship only in the shared
+asset.
 
 Skill-specific carve-outs that override or extend the shared rules are listed in the section
 synthesis rules below — most importantly: **the numeric `.yoy-change` cell stays
@@ -147,13 +143,12 @@ the scaffold for the final artifact. Do **not** copy it to the workspace and do
 For each company, fire the following calls in a single parallel batch (one message, many tool calls):
 
 ### Earnings call transcript (×4 per company) — primary source
-
-| Category label        | Keywords                                                                                                    |
-| --------------------- | ----------------------------------------------------------------------------------------------------------- |
-| Revenue & Price       | `Revenue, Net Sales, Sales, Volumes, Units sold, Price, Pricing`                                            |
-| Sector & Market       | `Industry, Sector Performance, Demand, Market Condition`                                                    |
+| Category label        | Keywords                                                                                            |
+|-----------------------|-----------------------------------------------------------------------------------------------------|
+| Revenue & Price       | `Revenue, Net Sales, Sales, Volumes, Units sold, Price, Pricing`                                    |
+| Sector & Market       | `Industry, Sector Performance, Demand, Market Condition`                                            |
 | Supply Chain & Region | `Supply Chain, Materials, Tariffs, Input, Logistics, Regional Performance, Geographic Condition, Geography` |
-| Guidance & Events     | `Business Outlook, Guidance, Revision, Corporate Events, Transaction`                                       |
+| Guidance & Events     | `Business Outlook, Guidance, Revision, Corporate Events, Transaction`                               |
 
 ### Web search fallback (only if `searchEntityEarningsCall` returns nothing usable)
 
@@ -192,9 +187,9 @@ The HTML file **must**:
   row labels, and element IDs exactly. Only the empty targets defined below are populated.
 - Be written in a single `create_file` call (no progressive `str_replace` edits, no
   multi-file split).
-  Render order of content inside the file follows the page top-to-bottom so the artifact is
-  human-readable as well as browser-renderable: cover/TOC fields first, then sections 1 → 13, then
-  sources.
+Render order of content inside the file follows the page top-to-bottom so the artifact is
+human-readable as well as browser-renderable: cover/TOC fields first, then sections 1 → 13, then
+sources.
 
 Use earnings call transcript content as the **sole** source for every section. Do not enrich
 with credit, sector outlook, ESG, peer, or news data — those tools are not permitted in this
@@ -206,7 +201,6 @@ URL-less fallback, and the rule that `n` matches the row position of the source
 `#ecs-sources` are defined in [skills/shared/citations/SKILL.md](../shared/citations/SKILL.md).
 
 ### Cover / header fields (write first)
-
 - `#ecs-report-date` — e.g. `April 4, 2026` (plain text)
 - `#ecs-footer-date` — same value (plain text)
 - `#ecs-company-count` — number of companies (plain text)
@@ -216,7 +210,6 @@ URL-less fallback, and the rule that `n` matches the row position of the source
   container (`<div class="cover-top has-cover-image">` and/or `<div class="cover-bottom has-cover-image">`).
   If you do not have images, leave the template as-is — the empty image strips will collapse
   automatically and the cover will render as a clean navy block with the accent bar.
-
 ### Section synthesis rules
 
 > **Per-section omit rule (applies to every section a–l).** If a company's earnings call
@@ -290,34 +283,33 @@ The `<strong class="subsection-title">…</strong>` subtitle **must be a single
 - `#ecs-sources` — end-of-document Citations rows. One `<div class="source-item">` per
   source, in `[1], [2], …` order, using the canonical row markup from the shared citations
   skill.
-
 ---
 
 ## Streaming protocol (element → content mapping)
 
-| Element ID                    | Content type                                                                                                                    |
-| ----------------------------- | ------------------------------------------------------------------------------------------------------------------------------- |
-| `#ecs-report-date`            | Plain text date                                                                                                                 |
-| `#ecs-footer-date`            | Plain text date (same value)                                                                                                    |
-| `#ecs-company-count`          | Plain text integer                                                                                                              |
-| `#ecs-company-chips`          | Sequence of `<span class="company-chip">…</span>`                                                                               |
-| `#ecs-cover-img-right`        | `<img>` element — set `src` attribute                                                                                           |
-| `#ecs-cover-img-bottom`       | `<img>` element — set `src` attribute                                                                                           |
-| `#ecs-executive-summary`      | One or more `<p>`; prose may carry inline citations                                                                             |
-| `#ecs-credit-badge`           | Single `<div class="credit-badge …">` element                                                                                   |
-| `#ecs-overall-credit`         | `<p>` intro + `<ul><li>` bullets; prose may carry inline citations                                                              |
-| `#ecs-tbody-revenue`          | `<tr>` rows with `.yoy-change` class; commentary cell may carry inline citations                                                |
-| `#ecs-tbody-volumes`          | `<tr>` rows with `.yoy-change` class; commentary cell may carry inline citations                                                |
-| `#ecs-tbody-prices`           | `<tr>` rows with `.yoy-change` class; commentary cell may carry inline citations                                                |
-| `#ecs-qualitative`            | `<strong class="subsection-title">` + `<p>` (×3); prose may carry inline citations                                              |
-| `#ecs-end-markets`            | `<strong class="subsection-title">` + `<ul><li>` (×4); bullets may carry inline citations                                       |
-| `#ecs-supply-chain`           | `<strong class="subsection-title">` + `<ul><li>` (×3); bullets may carry inline citations                                       |
-| `#ecs-geography`              | `<strong class="subsection-title">Region: Summary</strong>` + `<ul><li>` (×4 fixed regions); bullets may carry inline citations |
-| `#ecs-outlook`                | `<strong class="subsection-title">` + `<ul><li>` per company; bullets may carry inline citations                                |
-| `#ecs-revisions`              | `<strong class="subsection-title">` + `<ul><li>` per company; bullets may carry inline citations                                |
-| `#ecs-events`                 | `<strong class="subsection-title">` + `<ul><li>` per company; bullets may carry inline citations                                |
-| `#ecs-cite-a` … `#ecs-cite-l` | `<div class="section-citations">…</div>` (optional, see shared citations skill)                                                 |
-| `#ecs-sources`                | `<div class="source-item">` rows (see shared citations skill)                                                                   |
+| Element ID                       | Content type                                                       |
+|----------------------------------|--------------------------------------------------------------------|
+| `#ecs-report-date`               | Plain text date                                                    |
+| `#ecs-footer-date`               | Plain text date (same value)                                       |
+| `#ecs-company-count`             | Plain text integer                                                 |
+| `#ecs-company-chips`             | Sequence of `<span class="company-chip">…</span>`                  |
+| `#ecs-cover-img-right`           | `<img>` element — set `src` attribute                              |
+| `#ecs-cover-img-bottom`          | `<img>` element — set `src` attribute                              |
+| `#ecs-executive-summary`         | One or more `<p>`; prose may carry inline citations |
+| `#ecs-credit-badge`              | Single `<div class="credit-badge …">` element                      |
+| `#ecs-overall-credit`            | `<p>` intro + `<ul><li>` bullets; prose may carry inline citations |
+| `#ecs-tbody-revenue`             | `<tr>` rows with `.yoy-change` class; commentary cell may carry inline citations |
+| `#ecs-tbody-volumes`             | `<tr>` rows with `.yoy-change` class; commentary cell may carry inline citations |
+| `#ecs-tbody-prices`              | `<tr>` rows with `.yoy-change` class; commentary cell may carry inline citations |
+| `#ecs-qualitative`               | `<strong class="subsection-title">` + `<p>` (×3); prose may carry inline citations |
+| `#ecs-end-markets`               | `<strong class="subsection-title">` + `<ul><li>` (×4); bullets may carry inline citations |
+| `#ecs-supply-chain`              | `<strong class="subsection-title">` + `<ul><li>` (×3); bullets may carry inline citations |
+| `#ecs-geography`                 | `<strong class="subsection-title">Region: Summary</strong>` + `<ul><li>` (×4 fixed regions); bullets may carry inline citations |
+| `#ecs-outlook`                   | `<strong class="subsection-title">` + `<ul><li>` per company; bullets may carry inline citations |
+| `#ecs-revisions`                 | `<strong class="subsection-title">` + `<ul><li>` per company; bullets may carry inline citations |
+| `#ecs-events`                    | `<strong class="subsection-title">` + `<ul><li>` per company; bullets may carry inline citations |
+| `#ecs-cite-a` … `#ecs-cite-l`    | `<div class="section-citations">…</div>` (optional, see shared citations skill) |
+| `#ecs-sources`                   | `<div class="source-item">` rows (see shared citations skill)      |
 
 ### Reference HTML snippets
 
@@ -327,21 +319,7 @@ The `<strong class="subsection-title">…</strong>` subtitle **must be a single
 <tr>
   <td class="company-name">Company A</td>
   <td class="yoy-change up">+5.20%</td>
-  <td>
-    Revenue growth driven by pricing gains in North America
-    <a
-      href="https://example.com/transcript-q1-2026"
-      target="_blank"
-      class="cite-ref"
-      >[1]</a
-    >, partially offset by volume declines in EMEA
-    <a
-      href="https://example.com/credit-opinion-a"
-      target="_blank"
-      class="cite-ref"
-      >[2]</a
-    >.
-  </td>
+  <td>Revenue growth driven by pricing gains in North America <a href="https://example.com/transcript-q1-2026" target="_blank" class="cite-ref">[1]</a>, partially offset by volume declines in EMEA <a href="https://example.com/credit-opinion-a" target="_blank" class="cite-ref">[2]</a>.</td>
 </tr>
 ```
 
@@ -362,24 +340,8 @@ The `<strong class="subsection-title">…</strong>` subtitle **must be a single
 ```html
 <strong class="subsection-title">North America</strong>
 <ul>
-  <li>
-    Retail demand held up against tariff pass-through
-    <a
-      href="https://example.com/transcript-q1-2026"
-      target="_blank"
-      class="cite-ref"
-      >[1]</a
-    >.
-  </li>
-  <li>
-    Pricing remained disciplined across grocery and consumables
-    <a
-      href="https://example.com/sector-outlook"
-      target="_blank"
-      class="cite-ref"
-      >[3]</a
-    >.
-  </li>
+  <li>Retail demand held up against tariff pass-through <a href="https://example.com/transcript-q1-2026" target="_blank" class="cite-ref">[1]</a>.</li>
+  <li>Pricing remained disciplined across grocery and consumables <a href="https://example.com/sector-outlook" target="_blank" class="cite-ref">[3]</a>.</li>
 </ul>
 ```
 
@@ -391,21 +353,18 @@ blocks, and the end-of-document Citations block.
 ### Class-selection rules
 
 **`.yoy-change`** modifier (on the second `<td>` of YOY rows):
-
 - `up` — value starts with `+`, equals `Up`, or parses to a positive number
 - `down` — value starts with `-`, equals `Down`, or parses to a negative number
 - `flat` — value equals `Flat` or `0`
 - `na` — value is `Not specified`, `N/A`, empty, or `—`
-  **`.credit-badge`** modifier (for `#ecs-credit-badge`):
+**`.credit-badge`** modifier (for `#ecs-credit-badge`):
 - `positive`, `negative`, or `neutral`. Label text is the capitalized sentiment.
-
 ### Conventions
 
 - Use `<p>` for paragraphs, `<ul><li>` for bullets, `<strong class="subsection-title">…</strong>` for bold subheaders inside prose containers.
 - Emit inline citations per [skills/shared/citations/SKILL.md](../shared/citations/SKILL.md). Inline citations are the primary attribution mechanism — the optional per-section blocks (`#ecs-cite-a` … `#ecs-cite-l`) remain available as supplementary summary boxes. Numeric `.yoy-change` cells stay citation-free.
 - Do NOT include overall section titles — the template already has those.
 - `yoy_change` values: `+X.XX%` or `-X.XX%` if numeric, else `Up` / `Down` / `Flat` / `Not specified`.
-
 ---
 
 ## Step 5 — Tell the user
diff --git a/plugins/moody-s/skills/moody-s-earnings-brief/agents/openai.yaml b/plugins/moody-s/skills/moody-s-earnings-brief/agents/openai.yaml
index e24cb89b..c85857d7 100644
--- a/plugins/moody-s/skills/moody-s-earnings-brief/agents/openai.yaml
+++ b/plugins/moody-s/skills/moody-s-earnings-brief/agents/openai.yaml
@@ -1,4 +1,4 @@
 interface:
   display_name: "Moody's Earnings Brief"
-  short_description: "Produce a comprehensive Earnings Brief HTML report combining company overview, financials, peer comparison, credit profile, risks, and ESG"
-  default_prompt: "Use Moody's Earnings Brief skill to resolve the target entity, gather Moody's data across all sections, and generate a complete Earnings Brief HTML report."
+  short_description: "Produce an Earnings Call Summary HTML report for 2–5 companies, summarizing transcript insights across revenue, volumes, pricing, supply chain, guidance, and corporate events"
+  default_prompt: "Use Moody's Earnings Brief skill to resolve the target companies, gather their earnings call transcripts, and generate a complete Earnings Call Summary HTML report."
diff --git a/plugins/moody-s/skills/moody-s-earnings-brief/assets/cover_img_0.png b/plugins/moody-s/skills/moody-s-earnings-brief/assets/cover_img_0.png
new file mode 100644
index 00000000..330447da
Binary files /dev/null and b/plugins/moody-s/skills/moody-s-earnings-brief/assets/cover_img_0.png differ
diff --git a/plugins/moody-s/skills/moody-s-earnings-brief/assets/cover_img_1.png b/plugins/moody-s/skills/moody-s-earnings-brief/assets/cover_img_1.png
new file mode 100644
index 00000000..eab8e63e
Binary files /dev/null and b/plugins/moody-s/skills/moody-s-earnings-brief/assets/cover_img_1.png differ
diff --git a/plugins/moody-s/skills/moody-s-earnings-brief/assets/cover_img_2.png b/plugins/moody-s/skills/moody-s-earnings-brief/assets/cover_img_2.png
new file mode 100644
index 00000000..481aa989
Binary files /dev/null and b/plugins/moody-s/skills/moody-s-earnings-brief/assets/cover_img_2.png differ
diff --git a/plugins/moody-s/skills/moody-s-peer-analysis/SKILL.md b/plugins/moody-s/skills/moody-s-peer-analysis/SKILL.md
index cd4db77e..6763e586 100644
--- a/plugins/moody-s/skills/moody-s-peer-analysis/SKILL.md
+++ b/plugins/moody-s/skills/moody-s-peer-analysis/SKILL.md
@@ -14,39 +14,42 @@ Generates a professional HTML report (styled like a Moody's peer analysis PDF) f
 company and up to 3 credit peers. Data is pulled from multiple `Moodys MCP server` MCP tools and
 consolidated into a single HTML artifact covering peers comparison, ratings chart, and ESG.
 
-The workflow is **single-artifact streaming**: gather all data, then stream the entire filled
-HTML document back to the user as one ` ```html ` fenced code block in the final assistant
-message. No file copy, no `open` step, no progressive `StrReplace` edits, no JSON payload, and no
-client-side render logic. The fenced code block is the deliverable.
+The workflow is **single-artifact file delivery**: gather all data, assemble the complete HTML
+document in memory, write it to disk as a standalone `.html` file, and present it to the user for
+download. No fenced code blocks, no streaming partial sections, no JSON payload, no progressive
+edits. The downloaded file is the sole deliverable.
 
 > ## ⚠️ CRITICAL — NON-NEGOTIABLE OUTPUT CONTRACT
 >
-> **The LLM MUST stream the final report back as a single HTML artifact inside the assistant
-> response.** This is the only acceptable form of delivery for this skill. Specifically:
+> **The LLM MUST write the final report to disk and present it as a downloadable file.**
+> This is the only acceptable form of delivery for this skill. Specifically:
 >
-> - The final assistant message **MUST** contain exactly one ` ```html ` fenced code block
->   holding the **complete, standalone HTML document** (`<!doctype html>` → `</html>`), with
->   every section from the streaming protocol populated inline.
-> - The LLM **MUST NOT** write the report to a file on disk (no `Write`, no `cp` of the
->   template, no `StrReplace` into a working artifact, no `open` command).
-> - The LLM **MUST NOT** split the report across multiple code blocks, multiple messages,
->   partial snippets, or summaries.
-> - The LLM **MUST NOT** substitute prose, Markdown, JSON, attachments, or links for the
->   fenced HTML artifact. The artifact itself is the answer.
-> - If data gathering fails partially, still emit the single ` ```html ` artifact with
->   the best-available content and `"--"` placeholders for missing cells — never skip the
->   artifact.
+> - The LLM **MUST** write the complete, standalone HTML document (`<!doctype html>` → `</html>`)
+>   to `/mnt/user-data/outputs/{target_company_slug}_peer_analysis.html` and then call
+>   `present_files` to surface it to the user.
+> - The LLM **MUST NOT** print the raw HTML source in the chat — no ` ```html ` fenced code
+>   blocks, no partial snippets, no inline HTML in the assistant message.
+> - The LLM **MUST NOT** split the report across multiple files or multiple messages.
+> - The LLM **MUST NOT** substitute prose, Markdown, JSON, or links for the file download.
+> - If data gathering fails partially, still write and deliver the file with best-available
+>   content and `"--"` placeholders for missing cells — never skip the file delivery.
 >
 > Treat any other output shape as a hard failure of the skill.
 
-## Required MCP server
+## Required MCP servers
 
 `Moodys MCP server` — tools used: `findEntity`, `getEntityPeers`, `getEntityRatings`,
 `getEntityCreditOpinion`, `getEntityFinancials` (sections: Profile, Summary, RatingOutlook, FactorsLeadingToUpgrade,
 FactorsLeadingToDowngrade, CreditStrengths, CreditChallenges, ESGConsiderations,
 KeyIndicatorsTable, ScorecardTable), `getEntityEsg`, `getEntitySectorOutlook`
 
-If any of the tools required for a section do not exist, inform the user: One or more tools required for this section are not available under your current subscription. Unlock more of the expert insights, data, and analytics you trust. Get Link:https://www.moodys.com/web/en/us/capabilities/gen-ai/ai-ready-data.html with us to learn more. 
+`ProbabilityOfDefault - Prod - Moodys MCP - NEW URL` — PD tools:
+- `getEntityProbabilityOfDefault` — retrieve Current PD, 1-year PD, 5-year cumulative PD,
+  and implied rating for the target company AND each peer. Used to build the PD comparison
+  panel in Section 2. Set `pd_comparison` to `null` in the JSON payload if data is
+  unavailable for the target or fewer than 2 peers.
+
+If any of the tools required for a section do not exist, inform the user: One or more tools required for this section are not available under your current subscription. Unlock more of the expert insights, data, and analytics you trust. Get Link:https://www.moodys.com/web/en/us/capabilities/gen-ai/ai-ready-data.html with us to learn more.
 
 ## Bundled files
 
@@ -153,6 +156,14 @@ parallel batch** (one message, many tool calls):
 
 Call `getEntityCreditOpinion`,`getEntityFinancials` requesting these sections, about the key Indicators table this should be populated with most recent fiscal year available (e.g. 2025 FY) on `getEntityFinancials`. Not LTM allowed:
 
+> **CRITICAL — `getEntityFinancials` is the authoritative source for all quantitative data.**
+> Always call `getEntityFinancials` (with `excludeInterimData: true`) for every entity in the
+> peer set to obtain the most recent **fiscal year-end** figures. Do NOT rely solely on the
+> Key Indicators or Scorecard exhibits embedded in `getEntityCreditOpinion` — those may
+> reference an older period. After fetching financials, identify each entity's latest year
+> with a non-null Revenue value and use that year exclusively for Key Indicators charts,
+> Scorecard "Current" column, and the Revenue bar chart. LTM figures are **not permitted**.
+
 | Section parameter | Purpose |
 |---|---|
 | `Profile` | Company description for the Peers Table |
@@ -177,23 +188,37 @@ Call `getEntityEsg` — overall CIS classification plus E, S, G sub-scores.
 Call `getEntitySectorOutlook` for the target company's sector. Reuse for peers in the same
 sector.
 
+### Probability of Default (target + all peers)
+
+Call `getEntityProbabilityOfDefault` for the target company AND every resolved peer in a
+single parallel batch. Retrieve:
+- 1-year PD (%)
+- 5-year cumulative PD (%)
+- PD-implied rating
+
+Compute the **peer average** for 1-year PD and 5-year cumulative PD (exclude the target).
+Determine `pd_signal`:
+- `"red"` — target 1-year PD is >2× peer average
+- `"amber"` — target 1-year PD is 1.25–2× peer average
+- `"green"` — target 1-year PD is ≤1.25× peer average
+
+Build `pd_gap_1yr` label (e.g. `"+0.079pp above peer avg (3.5×)"`).
+
+Set `pd_comparison` to `null` if the PD tool is unavailable or data is missing for the target
+or fewer than 2 peers — the panel is suppressed automatically.
+
 Hold all results in context for Step 5 synthesis.
 
 ---
 
 ## Step 5 — Synthesize + emit the complete artifact
 
-After data is gathered, produce **one** final assistant message. The message contains:
+After data is gathered, assemble the complete filled HTML document and write it to disk, then
+present it to the user for download (see Step 7 for the file-naming and delivery contract).
 
-1. A one-line summary sentence (e.g. `Peer Analysis for {Target Company}:`).
-2. A single fenced ` ```html ` code block containing the **entire filled `template.html`
-   document** — with every element from the streaming protocol populated in place. No partial
-   documents, no separate code blocks per section.
+The assembled document **must**:
 
-The code block **must**:
-
-- Start at column 0 with ` ```html ` and end with a closing ` ``` ` on its own line.
-- Contain a complete, standalone HTML document (doctype → `</html>`) that renders without
+- Be a complete, standalone HTML document (doctype → `</html>`) that renders without
   external dependencies.
 - Preserve the template's `<head>` (CSS, fonts), section order, table skeletons, row labels, and
   element IDs exactly. Only the empty targets defined below are populated.
@@ -243,37 +268,111 @@ Write the 4 company header cells first:
 - `#pa-cd-col-2`, `#pa-cd-col-3`, `#pa-cd-col-4` — peer names in order (leave empty if fewer
   than 3 peers were resolved)
 
-Then write the 20 body cells (`#pa-cd-r{1..5}-c{1..4}`). Each cell is a substantive paragraph
+Then write the 16 body cells (`#pa-cd-r{1..4}-c{1..4}`). Each cell is a substantive paragraph
 with specific quantitative thresholds where available. Cells may carry inline citations:
 
 - Row 1 (`Upgrade factors`) — from `FactorsLeadingToUpgrade`
 - Row 2 (`Downgrade factors`) — from `FactorsLeadingToDowngrade`
 - Row 3 (`Credit strengths (qualitative)`) — from `CreditStrengths`
 - Row 4 (`Credit challenges (qualitative)`) — from `CreditChallenges`
-- Row 5 (`Quantitative support (most recent financials in context)`) — from
-  `KeyIndicatorsTable`; cite the most recent period's key metrics (debt/EBITDA, EBITA margin,
-  EBITA/interest, RCF/net debt, revenue)
+
+> **Row 5 ("Quantitative support") has been removed.** Do NOT emit a 5th row in the Credit
+> Drivers table. The pre-shaped table in `assets/template.html` has 5 row slots — leave
+> `#pa-cd-r5-c{1..4}` blank and suppress the row label for Row 5 by setting its `<th>` to
+> an empty string or by omitting it entirely if the template allows. Quantitative metrics
+> are covered by the Key Indicators bar charts immediately below.
 
 If a Credit Opinion section is missing for a company, write `Not available` in the corresponding
 cell rather than leaving it blank.
 
-**4) Key Indicators** — 4 rows × 11 columns, pre-shaped `<table>`
+**4) Key Indicators — Peer Bar Charts** → `#pa-ki-charts`
+
+Replace the static Key Indicators table with a set of **horizontal peer bar charts** matching
+the visual style from the sector-drift report. Emit a `<div id="pa-ki-charts">` block
+containing **five** `peer-bar-chart-outer` panels, one per metric:
+
+| Chart | Metric | Unit | Source | Note on "lower is better" |
+|-------|--------|------|--------|--------------------------|
+| 1 | FCF Margin % | `%` | (OCF − \|Capex\|) / Revenue | Higher = better |
+| 2 | Debt / EBITDA | `x` | Moody's pre-computed Key Indicators | Lower = better |
+| 3 | EBITDA Margin % | `%` | EBITDA / Revenue | Higher = better |
+| 4 | EBITA / Interest | `x` | Moody's pre-computed Key Indicators | Higher = better |
+| 5 | RCF / Net Debt | `%` | Moody's pre-computed Key Indicators | Higher = better |
+
+For each chart:
+- Show one bar per company (target first, then peers sorted by descending value, then AVERAGE last)
+- The target bar uses `background:#0066cc` (blue); AVERAGE bar uses `background:#2e7d32` (green)
+- Peers use `background:#7ca8d8` (steel blue) for positive values, `background:#e0a0a0` for negative
+- Draw an orange median line (`background: #e87722`) at the peer median position
+- **Show the fiscal year for each company's bar** by appending the FY label in parentheses to
+  the `pbc-name` element (e.g. `Target Co (FY2025)`, `Peer A (FY2024)`). This is required
+  because different entities may have different most-recent fiscal years. Use the same FY
+  year identified via `getEntityFinancials` for each entity.
+- Display the fiscal year reference at the top of the block (e.g. `FY2025, % of revenue | Orange line = peer median | Green bar = peer average`); if entities have mixed FYs, write `Most recent FY` instead of a single year
+- Include a note line at the bottom with: peer median, peer avg (n=X), company value, gap
+
+**Compute peer average and peer median** across all resolved peers (exclude the target).
+
+**FY period header**: show the most recent common FY year at the top of `#pa-ki-charts` in
+a `<div class="ki-fy-header">` (e.g. `KEY INDICATORS — FY2025 | Company vs. Peers`).
 
-For each company row `r ∈ {1..4}` (target = row 1), write the 11 cells:
-- `#pa-ki-r{r}-company` — company name
-- `#pa-ki-r{r}-v1` — Period (e.g. `2025 FY`)
-- `#pa-ki-r{r}-v2` — Revenue (USD)
-- `#pa-ki-r{r}-v3` — Total Debt (USD)
-- `#pa-ki-r{r}-v4` — EBITA (USD)
-- `#pa-ki-r{r}-v5` — EBITA Margin
-- `#pa-ki-r{r}-v6` — EBITDA (USD)
-- `#pa-ki-r{r}-v7` — Debt/EBITDA
-- `#pa-ki-r{r}-v8` — EBITA/Interest
-- `#pa-ki-r{r}-v9` — Net LT Debt (USD)
-- `#pa-ki-r{r}-v10` — RCF/Net Debt
+**Reference snippet** for one chart panel (emit five of these inside `#pa-ki-charts`):
 
-Source all values from `getEntityFinancials`. If fewer than 3 peers are
-resolved, leave the unused row's cells empty.
+```html
+<div id="pa-ki-charts">
+  <div class="ki-fy-header">KEY INDICATORS — FY2025 | Company vs. Peers</div>
+
+  <div class="peer-bar-chart-outer">
+    <div class="pbc-title">FCF MARGIN — COMPANY VS. PEERS</div>
+    <div class="pbc-subtitle">FY2025, % of revenue | Orange line = peer median | Green bar = peer average</div>
+    <div class="pbc-rows">
+      <!-- Target (subject) row -->
+      <div class="pbc-row">
+        <div class="pbc-name subject" title="Target Co">Target Co</div>
+        <div class="pbc-track">
+          <div class="pbc-median-line" style="left:52.3%"><div class="pbc-median-tip">Median 8.4%</div></div>
+          <div class="pbc-fill subj-neg" style="left:0;width:12.1%"></div>
+        </div>
+        <div class="pbc-val neg">-2.1%</div>
+      </div>
+      <!-- Peer rows (sorted descending by value) -->
+      <div class="pbc-row">
+        <div class="pbc-name" title="Peer A">Peer A</div>
+        <div class="pbc-track">
+          <div class="pbc-median-line" style="left:52.3%"></div>
+          <div class="pbc-fill peer-pos" style="left:0;width:100%"></div>
+        </div>
+        <div class="pbc-val pos">12.3%</div>
+      </div>
+      <!-- AVERAGE bar (last, separated by dashed border) -->
+      <div class="pbc-row" style="margin-top:6px;padding-top:6px;border-top:1px dashed #d0d7e3">
+        <div class="pbc-name average">AVERAGE</div>
+        <div class="pbc-track">
+          <div class="pbc-fill avg-bar" style="left:0;width:70.7%"></div>
+        </div>
+        <div class="pbc-val avg-val">8.7%</div>
+      </div>
+    </div>
+    <div class="pbc-note">Peer median 8.42% | Peer avg 8.70% (n=6, Moody's). Company: -2.10%. Gap vs. median: -10.52pp.</div>
+  </div>
+
+  <!-- Repeat for Debt/EBITDA, EBITDA Margin %, EBITA/Interest, RCF/Net Debt -->
+</div>
+```
+
+**Bar width calculation** — for charts with only positive values:
+- Find `maxAbs` = max of all absolute values across company + peers
+- `fillW` = `(value / maxAbs * 88).toFixed(1) + '%'`
+
+For charts that could include negative values (FCF Margin):
+- Use a zero-anchored track: `zeroPct = 44`, positive bars grow right, negative bars grow left
+- `fillW = (|value| / maxAbs * 44).toFixed(1) + '%'`
+- `fillL` for negative: `(zeroPct - |value|/maxAbs*44).toFixed(1) + '%'`
+
+**Median line position**: `medPct = (medV / maxAbs * 88).toFixed(1) + '%'` (positive-only charts)
+
+Source all values from `getEntityFinancials` (most recent FY — same period used in scorecards).
+If a value is missing, omit the bar row for that company.
 
 **5) Scorecards** — 7 rows × 8 sub-columns (4 companies × Current/Forward), pre-shaped `<table>`
 
@@ -282,9 +381,41 @@ Write the 4 top-level company headers first:
 - `#pa-sc-col-2`, `#pa-sc-col-3`, `#pa-sc-col-4` — peer names
 
 Write the 8 sub-header labels:
-- `#pa-sc-col-{1..4}-curr` — current-period label (e.g. `Current 2025 FY`)
+- `#pa-sc-col-{1..4}-curr` — current-period label that **must exactly match the most-recent
+  FY period used for that entity in the Key Indicators bar charts** (e.g. if Apple's KI data
+  is from FY2025, write `FY2025`; if Microsoft's is from FY2024, write `FY2024`). Each
+  entity's label is set independently — do not force a single common year across all columns.
 - `#pa-sc-col-{1..4}-fwd` — forward-period label (e.g. `Forward`)
 
+> **FY LABEL SYNCHRONY — STRICTLY ENFORCED (Sections 4 and 5):**
+> The fiscal year label shown in Section 4 (Key Indicators bar charts) for each entity
+> **MUST be identical** to the label shown in the `#pa-sc-col-{n}-curr` sub-header for that
+> same entity in Section 5 (Scorecards). This is a hard consistency requirement:
+>
+> - Determine the most-recent non-null full fiscal year-end for each entity from
+>   `getEntityFinancials` (with `excludeInterimData: true`). Label it as `FY{YYYY}`
+>   (e.g. `FY2025`, `FY2024`). **Never use "LTM", "TTM", "Current", or any period
+>   label other than the specific `FY{YYYY}` form.**
+> - Write that exact `FY{YYYY}` string in:
+>   1. The `pbc-name` element of each company's row inside every Section 4 bar chart
+>      (appended in parentheses, e.g. `Apple Inc. (FY2025)`), AND in the
+>      `<div class="ki-fy-header">` block label.
+>   2. The `#pa-sc-col-{n}-curr` sub-header cell for that entity in Section 5.
+>   3. The data values themselves — all numbers in Section 4 and Section 5 Current cells
+>      must be sourced from the same `FY{YYYY}` period.
+> - If two entities have different most-recent FYs, each gets its own correct label
+>   independently. The header `<div class="ki-fy-header">` should then read
+>   `KEY INDICATORS — Most Recent FY | Company vs. Peers`.
+> - This synchrony check is mandatory before emitting the final HTML. Mismatched labels
+>   between Section 4 and Section 5 are a hard failure of the skill.
+
+> **`~` (tilde) prefix rule — STRICT:** The `~` prefix (e.g. `~2.5x`, `~15%`) means
+> *approximately estimated / forecast* and must **only** appear in **Forward** (`-fwd`) cells.
+> It must **never** appear in **Current** (`-curr`) cells. Current-period cells contain actual
+> reported or Moody's-adjusted figures sourced from `getEntityFinancials`; use the precise
+> value as-is (e.g. `2.3x`, `14.8%`). If the exact current figure is unavailable, write `N/A`
+> — do not substitute a tilde-prefixed estimate.
+
 Then write the 56 body cells (`#pa-sc-r{1..7}-c{1..4}-curr` and `-fwd`) for the 7 rows:
 - Row 1 — Scale: Revenue/Sales (USD bn)
 - Row 2 — Profitability: EBIT(A) margin
@@ -308,7 +439,21 @@ may carry inline citations:
 - Paragraph 3: Practical positioning — frame the target's "path to stand out" by linking back
   to the upgrade/downgrade framework and key risk factors.
 
-### Section 2 — Ratings Chart
+### Section 2 — Ratings Chart (± Probability of Default)
+
+> **SECTION 2 TITLE RULE — STRICTLY ENFORCED:**
+> The heading for Section 2 (in both the TOC `<li>` and the `<div class="section-heading">`)
+> must be set **conditionally** at emit time:
+>
+> - If PD data is available (i.e. `pd_comparison` is NOT `null`) →
+>   use **"Section 2. Ratings Chart & Probability of Default"**
+> - If PD data is unavailable (i.e. `pd_comparison` is `null`, the MCP server is unreachable,
+>   or data is missing for the target or fewer than 2 peers) →
+>   use **"Section 2. Ratings Chart"**
+>
+> The template ships with the longer title as a placeholder; overwrite it in both locations
+> (`#sec2 .section-heading` and the matching `<li>` in `<ul class="toc-list">`) whenever
+> PD data is absent. Never show "& Probability of Default" in the title if there is no PD panel.
 
 **Chart** → `#pa-ratings-chart`
 
@@ -317,6 +462,64 @@ plots the last 5 rating actions for each company as a line chart. X-axis uses se
 (`Rating 1` … `Rating 5`, most recent first); Y-axis maps Moody's rating symbols to a numeric
 scale.
 
+**PD Panel** → `#pa-pd-panel`
+
+Immediately after the ratings chart, emit the Probability of Default panel as a **horizontal
+bar chart** using the same `.peer-bar-chart-outer` style as the Key Indicators charts in
+Section 1. If PD data is unavailable, leave `#pa-pd-panel` empty (the element collapses).
+
+Emit one `peer-bar-chart-outer` block showing the **1-year PD (%)** for each company. Use the
+same CSS classes and bar-width formula as the KI charts (`pbc-fill`, `pbc-track`, `pbc-rows`,
+etc.). Include a note row at the bottom with peer median, peer average, and the gap signal.
+
+Signal colouring for the target bar:
+- `pbc-fill subj-neg` (red) — target 1-year PD > 2× peer average (`pd_signal: "red"`)
+- `pbc-fill subj-pos` (blue) — all other cases
+
+**Reference snippet** (emit this block directly into `#pa-pd-panel`):
+
+```html
+<div class="peer-bar-chart-outer">
+  <div class="pbc-title">PROBABILITY OF DEFAULT — 1-YEAR PD (%)</div>
+  <div class="pbc-subtitle">Moody's PD model | Orange line = peer median | Green bar = peer average</div>
+  <div class="pbc-rows">
+    <!-- Target (subject) row — blue or red depending on pd_signal -->
+    <div class="pbc-row">
+      <div class="pbc-name subject" title="Target Co">Target Co</div>
+      <div class="pbc-track">
+        <div class="pbc-median-line" style="left:38.5%"><div class="pbc-median-tip">Median 0.032%</div></div>
+        <div class="pbc-fill subj-neg" style="left:0;width:100%"></div>
+      </div>
+      <div class="pbc-val neg">0.111%</div>
+    </div>
+    <!-- Peer rows (sorted descending by pd_1yr) -->
+    <div class="pbc-row">
+      <div class="pbc-name" title="Peer A">Peer A</div>
+      <div class="pbc-track">
+        <div class="pbc-median-line" style="left:38.5%"></div>
+        <div class="pbc-fill peer-pos" style="left:0;width:54.1%"></div>
+      </div>
+      <div class="pbc-val pos">0.060%</div>
+    </div>
+    <!-- AVERAGE bar (last, separated by dashed border) -->
+    <div class="pbc-row" style="margin-top:6px;padding-top:6px;border-top:1px dashed #d0d7e3">
+      <div class="pbc-name average">AVERAGE</div>
+      <div class="pbc-track">
+        <div class="pbc-fill avg-bar" style="left:0;width:28.8%"></div>
+      </div>
+      <div class="pbc-val avg-val">0.032%</div>
+    </div>
+  </div>
+  <div class="pbc-note">Peer median 0.032% | Peer avg 0.032% (n=3). Target: 0.111%. Gap: +0.079pp above peer avg (3.5×). PD-implied rating: A1.</div>
+  <div class="pbc-note" style="margin-top:4px;font-style:italic">One-sentence PD narrative goes here.</div>
+</div>
+```
+
+**Bar width calculation:** `fillW = (pd_1yr / maxPD * 88).toFixed(1) + '%'` where `maxPD` is
+the maximum 1-year PD across all companies including the target.
+
+**Median line position:** `medPct = (medianPD / maxPD * 88).toFixed(1) + '%'`
+
 **Analysis** → `#pa-ratings-analysis`
 
 One `<p>` paragraph on rating trajectories — which companies show improvement, deterioration,
@@ -364,14 +567,14 @@ Two `<p>` paragraphs. Paragraphs may carry inline citations:
 | `<tbody id="pa-peers-table">`                        | `<tr>` rows (company name + description) with inline citations         |
 | `<tbody id="pa-peers-rating">`                       | `<tr>` rows with `<span class="outlook-badge …">`                      |
 | `#pa-cd-col-1` … `#pa-cd-col-4`                      | Plain text (company names)                                             |
-| `#pa-cd-r{1..5}-c{1..4}`                             | Prose cell content (short `<p>` or plain text) with inline citations   |
-| `#pa-ki-r{1..4}-company`                             | Plain text (company name)                                              |
-| `#pa-ki-r{1..4}-v{1..10}`                            | Plain text (period / metric value)                                     |
+| `#pa-cd-r{1..4}-c{1..4}`                             | Prose cell content (short `<p>` or plain text) with inline citations   |
+| `#pa-ki-charts`                                      | Five `.peer-bar-chart-outer` panels (FCF Margin, Debt/EBITDA, EBITDA Margin, EBITA/Interest, RCF/Net Debt) |
 | `#pa-sc-col-1` … `#pa-sc-col-4`                      | Plain text (company names, span 2 sub-columns)                         |
 | `#pa-sc-col-{1..4}-curr`, `#pa-sc-col-{1..4}-fwd`    | Plain text sub-header labels (Current / Forward period)                |
 | `#pa-sc-r{1..7}-c{1..4}-curr`, `…-fwd`               | Plain text scorecard cell values                                       |
 | `#pa-conclusion`                                     | Three `<p>` paragraphs with inline citations                           |
 | `#pa-ratings-chart`                                  | Single `<div class="chart-container">…<svg>…</svg>…</div>` block       |
+| `#pa-pd-panel`                                       | PD comparison panel — horizontal bar chart using `peer-bar-chart-outer` style |
 | `#pa-ratings-analysis`                               | One `<p>` paragraph with inline citations                              |
 | `<tbody id="pa-esg-table">`                          | `<tr>` rows (company + overall + E + S + G)                            |
 | `#pa-esg-analysis`                                   | Two `<p>` paragraphs with inline citations                             |
@@ -589,20 +792,19 @@ Per point `(i, v)` where `i` is the x-index (0-based) and `v` is the rating valu
 
 ---
 
-## Step 6 — Tell the user
 
-After the ` ```html ` code block, add a single short sentence confirming the artifact is
-complete (e.g. `Artifact rendered above — the report is fully self-contained HTML.`). Do not
-write the artifact to disk and do not suggest shell commands. The code block itself is the
-deliverable.
+
+
+
+
+
 
 ---
 
 ## Tips
 
 - Run ALL data-gathering tool calls in a single parallel batch (one message, many tool calls).
-- Emit the final HTML as one `html` fenced code block — do not stream partial sections, do not
-  split across multiple messages, do not write to disk.
+- Write the final HTML to disk as a single `.html` file and deliver it via `present_files` — do not print the HTML source in the chat, do not split across multiple files.
 - The target company always appears as the first column / first row in every table and
   comparison.
 - The conclusion must orient around the target company — how it differentiates from its peers.
@@ -620,6 +822,16 @@ deliverable.
 - Inline citations follow the shared citations skill — read
   [skills/shared/citations/SKILL.md](../shared/citations/SKILL.md) before authoring any `[n]`
   reference or the Citations block.
+- **PD panel**: write the fully rendered HTML directly into `#pa-pd-panel` using the CSS classes
+  already in the template (`.pd-panel`, `.pd-stats`, `.pbc-rows`, etc.). The template's `<script>`
+  block handles the `data-pd` attribute approach as a fallback, but writing innerHTML directly is
+  preferred in the static-artifact delivery mode. If PD data is unavailable, leave `#pa-pd-panel`
+  empty (the section collapses automatically).
+- **Key Indicators bar charts**: write the complete `#pa-ki-charts` block directly — five
+  `.peer-bar-chart-outer` panels. Use the CSS classes `.pbc-rows`, `.pbc-row`, `.pbc-name`,
+  `.pbc-track`, `.pbc-fill`, `.pbc-val`, `.pbc-median-line`, `.pbc-median-tip` already defined
+  in the template. The FY period header goes in `<div class="ki-fy-header">` as the first child
+  of `#pa-ki-charts`.
 
 ---
 
@@ -666,7 +878,11 @@ differs from another's.
    Scorecard, regardless of what year the credit opinion's exhibit uses.
 3. If different entities have different most-recent years (e.g. one entity has 2025 data
    while another only has 2024), use each entity's own most recent year independently and
-   label the `#pa-ki-r{n}-v1` cell accordingly (e.g. `2025 FY` vs `2024 FY`).
+   label the `#pa-ki-r{n}-v1` cell accordingly (e.g. `2025 FY` vs `2024 FY`). The
+   `#pa-sc-col-{n}-curr` sub-header for that entity's Scorecard column **must show the same
+   FY label** used in the Key Indicators for that entity — this is the consistency requirement.
+   Never use a generic label like `Current` alone; always include the specific year
+   (e.g. `FY2025`, `FY2024`).
 
 ### Handling distorted EBITDA years
 
@@ -716,57 +932,71 @@ The target card carries the extra class `target-card`.
 
 ---
 
-### Visual 2 — Financial Comparison Bar Charts (`#pa-fi-charts`)
+### Visual 2 — Revenue Comparison Bar Chart (`#pa-fi-charts`)
 
-Placed **immediately after** the Key Indicators table (sub-heading 4), before sub-heading 5
-(Scorecards). Emit a two-column grid (`div.fi-chart-grid`) containing exactly **two**
-`div.fi-chart-box` blocks:
+Placed **immediately after** `#pa-ki-charts`, with **no separate sub-heading** between them.
+Both `#pa-ki-charts` and `#pa-fi-charts` appear under the same section 4 heading (`4) Key Indicators — Peer Bar Charts`). Do **not** emit a `<div class="sub-heading">` before `#pa-fi-charts`. Emit a **single** `peer-bar-chart-outer` panel for **Revenue** only, using
+exactly the same CSS classes and bar-width formula as the Key Indicators charts.
 
-1. **Revenue** — horizontal bar chart comparing Revenue (USD) for all companies
-2. **Debt/EBITDA** — horizontal bar chart comparing Debt/EBITDA ratio for all companies
+> **Note:** Debt/EBITDA is already covered in the Key Indicators charts — do NOT add a
+> second Debt/EBITDA chart here. Revenue is the only metric shown in `#pa-fi-charts`.
 
-For each chart, the widths are proportional: the company with the largest value gets a
-`fi-bar-fill` width of `100%`; all others are scaled relative to that maximum.
+**Source:** `getEntityFinancials` (same most-recent FY used in Key Indicators).
+**Format:** `$XB` (e.g. `$180.2B`); if missing write `N/A` and omit the bar.
 
-**Colour palette** (same as the ratings chart): target = `#0066cc`, peer 1 = `#e6550d`,
-peer 2 = `#1a7a4a`, peer 3 = `#8b5cf6`. Assign colours in the same company order as all
-other tables (target first).
+**Bar width calculation:** `fillW = (revenue / maxRevenue * 88).toFixed(1) + '%'` where
+`maxRevenue` is the maximum revenue across all companies **including** the target and the peer
+average bar.
 
-**Values:** source from `getEntityFinancials` (same figures used in Key Indicators).
-Format Revenue as `$XB` or `$XM`; Debt/EBITDA as `X.Xx`.
+**AVERAGE bar — REQUIRED:** Compute the **peer average revenue** (arithmetic mean of all
+resolved peers, excluding the target). Add an AVERAGE bar as the last row, separated by a
+dashed top border, using class `pbc-fill avg-bar` (green) and `pbc-name average`. The AVERAGE
+bar width uses the same formula: `(peerAvgRevenue / maxRevenue * 88).toFixed(1) + '%'`. If
+`peerAvgRevenue` exceeds `maxRevenue`, recalculate `maxRevenue` to include it.
 
-If a value is missing or `N/A`, set `fi-bar-fill` width to `0%` and display `N/A` in
-`.fi-bar-value`.
+No median line needed for Revenue.
+Target bar uses `pbc-fill subj-pos` (blue). Peer bars use `pbc-fill peer-pos`. Average bar
+uses `pbc-fill avg-bar` (green), label class `pbc-name average`, value class `pbc-val avg-val`.
 
-**Reference snippet** for one chart box (repeat once for Revenue, once for Debt/EBITDA):
+**Reference snippet:**
 
 ```html
-<div class="fi-chart-box">
-  <div class="fi-chart-title">Revenue (USD)</div>
-  <div class="fi-bar-row">
-    <span class="fi-bar-label">Target Co</span>
-    <div class="fi-bar-track"><div class="fi-bar-fill" style="width:100%;background:#0066cc;"></div></div>
-    <span class="fi-bar-value">$180B</span>
-  </div>
-  <div class="fi-bar-row">
-    <span class="fi-bar-label">Peer One</span>
-    <div class="fi-bar-track"><div class="fi-bar-fill" style="width:34%;background:#e6550d;"></div></div>
-    <span class="fi-bar-value">$61B</span>
-  </div>
-  <div class="fi-bar-row">
-    <span class="fi-bar-label">Peer Two</span>
-    <div class="fi-bar-track"><div class="fi-bar-fill" style="width:13%;background:#1a7a4a;"></div></div>
-    <span class="fi-bar-value">$23B</span>
-  </div>
-  <div class="fi-bar-row">
-    <span class="fi-bar-label">Peer Three</span>
-    <div class="fi-bar-track"><div class="fi-bar-fill" style="width:20%;background:#8b5cf6;"></div></div>
-    <span class="fi-bar-value">$36B</span>
+<div id="pa-fi-charts">
+  <div class="peer-bar-chart-outer">
+    <div class="pbc-title">REVENUE — COMPANY VS. PEERS</div>
+    <div class="pbc-subtitle">FY2025, USD billions</div>
+    <div class="pbc-rows">
+      <div class="pbc-row">
+        <div class="pbc-name subject" title="Target Co">Target Co</div>
+        <div class="pbc-track">
+          <div class="pbc-fill subj-pos" style="left:0;width:88%"></div>
+        </div>
+        <div class="pbc-val pos">$180.2B</div>
+      </div>
+      <div class="pbc-row">
+        <div class="pbc-name" title="Peer One">Peer One</div>
+        <div class="pbc-track">
+          <div class="pbc-fill peer-pos" style="left:0;width:29.7%"></div>
+        </div>
+        <div class="pbc-val pos">$61.0B</div>
+      </div>
+      <!-- Repeat for all resolved peers -->
+      <!-- AVERAGE bar — always last, separated by dashed border -->
+      <div class="pbc-row" style="margin-top:6px;padding-top:6px;border-top:1px dashed #d0d7e3">
+        <div class="pbc-name average">AVERAGE</div>
+        <div class="pbc-track">
+          <div class="pbc-fill avg-bar" style="left:0;width:XX.X%"></div>
+        </div>
+        <div class="pbc-val avg-val">$XX.XB</div>
+      </div>
+    </div>
+    <div class="pbc-note">Revenue sourced from Moody's financials (most recent FY). Peer avg (n=3): $XX.XB.</div>
   </div>
 </div>
 ```
 
-Omit bar rows for companies that were not resolved (fewer than 3 peers).
+Omit bar rows for companies that were not resolved (fewer than 3 peers). Always include the
+AVERAGE bar as long as at least one peer is resolved.
 
 ---
 
@@ -822,8 +1052,10 @@ If the score is `N/A` or missing, omit the colour class (plain white cell).
 | Element ID / selector          | Content                                                                 |
 |-------------------------------|-------------------------------------------------------------------------|
 | `#pa-rating-cards`            | `.rating-card` divs — one per company (target first)                   |
-| `#pa-fi-charts`               | Two `.fi-chart-box` divs inside `.fi-chart-grid` (Revenue, Debt/EBITDA)|
+| `#pa-ki-charts`               | Five `.peer-bar-chart-outer` bar chart panels (Key Indicators as bar charts) |
+| `#pa-fi-charts`               | Single `.peer-bar-chart-outer` panel (Revenue only)                    |
 | `#pa-esg-heatmap`             | `.esg-heatmap` table with colour-coded CIS / E / S / G scores          |
+| `#pa-pd-panel`                | PD comparison panel — rendered client-side from `pd_comparison` JSON   |
 
 ---
 
@@ -838,14 +1070,23 @@ Emit sub-sections in this exact order:
    Visual 1. **Do NOT emit the `<tbody id="pa-peers-rating">` table at all.** The cards are the
    sole representation of peer ratings in Section 1.
 3. **Credit Drivers table** — unchanged, as defined above.
-4. **Key Indicators table** + **Financial Comparison Bar Charts** (`#pa-fi-charts`) — unchanged,
-   both emitted as defined above (table first, then the two bar-chart boxes).
+4. **Key Indicators Bar Charts** (`#pa-ki-charts`) — five horizontal bar chart panels (FCF
+   Margin, Debt/EBITDA, EBITDA Margin, EBITA/Interest, RCF/Net Debt). **Do NOT emit the
+   `<table id="pa-key-indicators">` at all.** The bar charts are the sole representation of
+   key indicator data. Immediately after `#pa-ki-charts` (with **no separate sub-heading**),
+   emit the **Revenue Bar Chart** (`#pa-fi-charts`) — a single `peer-bar-chart-outer` panel
+   showing Revenue only. `#pa-fi-charts` is visually part of section 4; it must **not** be
+   preceded by its own `<div class="sub-heading">` element. **Do NOT add a Debt/EBITDA chart
+   here** — it is already present in the KI charts above.
 5. **Scorecard table** — unchanged, as defined above.
 6. **Conclusion** (`#pa-conclusion`) — unchanged.
 
 The `<table>` element whose `<tbody>` carries `id="pa-peers-rating"` must be **omitted entirely**
 from the emitted HTML. Do not render it, do not hide it with CSS, do not leave an empty wrapper.
 
+The `<table id="pa-key-indicators">` must also be **omitted entirely**. The bar charts in
+`#pa-ki-charts` replace it completely.
+
 ---
 
 ### Section 2 — Enhanced Ratings Chart
@@ -967,4 +1208,5 @@ readable, colour-coded format; the plain table is redundant and must be omitted.
 | Element | Reason |
 |---|---|
 | `<table>` wrapping `<tbody id="pa-peers-rating">` | Replaced by rating cards |
+| `<table id="pa-key-indicators">` | Replaced by peer bar charts (`#pa-ki-charts`) |
 | `<table>` wrapping `<tbody id="pa-esg-table">` | Replaced by ESG heatmap |
diff --git a/plugins/moody-s/skills/moody-s-rating-analysis/SKILL.md b/plugins/moody-s/skills/moody-s-rating-analysis/SKILL.md
index 5d6eaf6a..ab1a43cc 100644
--- a/plugins/moody-s/skills/moody-s-rating-analysis/SKILL.md
+++ b/plugins/moody-s/skills/moody-s-rating-analysis/SKILL.md
@@ -1,37 +1,36 @@
 ---
 name: moody-s-rating-analysis
 description: >
-  Produce a Rating Pitch Deck for a company using Moody's GenAI MCP tools, delivered as an
-  editable PowerPoint (.pptx) saved to disk. Use this skill whenever the user asks to create
-  a rating pitch, rating pitch deck, credit pitch, rating presentation, or rating pitch
-  report. Also trigger when they ask for a comprehensive credit overview combining sector
-  analysis, company financials, SWOT, peer comparison, and ESG into a single deck or
-  presentation. Trigger even if they just name a company and say "pitch deck", "rating deck",
-  or "credit deck".
+  Produce a Rating Pitch Report for a company using Moody's GenAI MCP tools, delivered as a
+  self-contained HTML file saved to disk. Use this skill whenever the user asks to create
+  a rating pitch, rating pitch deck, credit pitch, rating presentation, rating pitch
+  report, or rating HTML report. Also trigger when they ask for a comprehensive credit
+  overview combining sector analysis, company financials, SWOT, peer comparison, and ESG
+  into a single report or presentation. Trigger even if they just name a company and say
+  "pitch deck", "rating deck", "credit deck", or "rating report".
 ---
 
 # Rating Pitch Skill
 
-Generates a Moody's Rating Pitch Deck as an editable PowerPoint (.pptx) from a single MCP
-data pass. The Python builder (`scripts/build_pptx.py`) opens the official Moody's Corp 2026
-PowerPoint template, removes its demo slides, and populates the template's built-in layouts
-(Cover, Agenda, Dividers, 1/2/3-column content, Back Cover, Disclaimer) with the resolved
-payload data. The builder produces native, editable PowerPoint charts and tables end-to-end —
-no HTML preview, no in-chat artifact.
+Generates a Moody's Rating Pitch Report as a self-contained HTML file from a single MCP
+data pass. The Python builder (`scripts/build_html.py`) takes the resolved payload JSON and
+produces a single `.html` file containing all sections with inline Chart.js charts, styled
+tables, and bullet lists using the Moody's brand palette — no external dependencies beyond
+a browser to open it.
 
 > ## ⚠️ CRITICAL — NON-NEGOTIABLE OUTPUT CONTRACT
 >
-> Every run of this skill MUST produce an editable `.pptx`. Specifically:
+> Every run of this skill MUST produce a self-contained `.html` report. Specifically:
 >
 > - The skill **MUST** save the resolved `payload.json` to
->   `~/Desktop/rating-pitch/<company>-<YYYYMMDD-HHMMSS>/` and run `scripts/build_pptx.py`
->   against it to produce the editable `rating_pitch.pptx` alongside it.
-> - The LLM **MUST NOT** stream the deck content as inline HTML, Markdown, JSON dumps, or
->   any other in-chat artifact in lieu of building the `.pptx`. The `.pptx` itself is the
->   deliverable.
+>   `~/Desktop/rating-pitch/<company>-<YYYYMMDD-HHMMSS>/` and run `scripts/build_html.py`
+>   against it to produce the `rating_pitch.html` alongside it.
+> - The LLM **MUST NOT** stream the report content as inline Markdown, JSON dumps, or
+>   any other in-chat artifact in lieu of building the `.html` file. The `.html` file itself
+>   is the deliverable.
 > - The final assistant message **MUST** point the user at the full path to the generated
->   `rating_pitch.pptx` so they can open it.
-> - If data gathering fails partially, still build the `.pptx` from the partial payload
+>   `rating_pitch.html` so they can open it in their browser.
+> - If data gathering fails partially, still build the `.html` from the partial payload
 >   using `"--"` placeholders for missing values — never skip the build.
 >
 > Treat any other output shape as a hard failure of the skill.
@@ -52,11 +51,9 @@ If any of the tools required for a section do not exist, inform the user: One or
 
 ## Bundled files
 
-- `scripts/build_pptx.py` — the deck builder. Takes a JSON payload and emits a `.pptx`.
-- `scripts/requirements.txt` — Python dependencies (`python-pptx`). Unchanged.
-- `assets/Moody_Corp_Template.pptx` — official Moody's 2026 Corp PowerPoint template.
-  The builder opens this file, clears its demo slides, and populates its layouts.
-  **Do not modify.**
+- `scripts/build_html.py` — the report builder. Takes a JSON payload and emits a `.html`.
+  Uses only the Python standard library; no pip installs required.
+- `scripts/requirements.txt` — no additional Python dependencies needed.
 - `assets/sample_payload.json` — reference payload showing every field populated. Read this
   if you're ever unsure what a field should look like.
 
@@ -117,13 +114,13 @@ from the data and carry it through to `peer_financials.rows[].period` and
 
 Build a single in-memory **resolved payload** that matches the JSON shape in the **Payload
 schema** section below (a reference copy lives at `assets/sample_payload.json`). This
-payload drives the `.pptx` build (Step 4) — fill it completely before moving on.
+payload drives the .html build (Step 4) — fill it completely before moving on.
 
 Content rules for each section:
 
 > **`commentary` type rule — applies to every section without exception:**
 > All `commentary` fields in the payload MUST be a **JSON array of strings** — never a
-> bare string. A bare string passed to the `.pptx` builder is iterated character-by-character,
+> bare string. A bare string passed to the .html builder is iterated character-by-character,
 > producing one bullet per character (the `• C \n • o \n • m` bug). Always write:
 > `"commentary": ["Sentence one.", "Sentence two."]` — even for a single sentence.
 
@@ -177,34 +174,58 @@ Content rules for each section:
   the per-bar label makes the comparison transparent.
 - **peer_scatter** — two scatter series (`margin_vs_leverage`, `fcf_vs_rcf`), each a list
   of `{company, x, y}` points. Drop extreme outliers that would distort the axes.
+  > **Scatter chart rendering notes:**
+  > - Each company is rendered as a **separate series** so it gets its own distinct brand colour
+  >   (Blue → company 0, Pink → 1, Teal → 2, Gold → 3, Mid Blue → 4, Purple → 5).
+  > - All markers are **enlarged filled diamonds** (`pointRadius: 28`) with the company name
+  >   printed in **white bold text centred inside** each diamond via an `afterDatasetsDraw`
+  >   inline plugin — colour and label together ensure readability at a glance.
+  > - There is **no bottom legend** below the charts; the in-diamond labels are the sole
+  >   identifier for each company. Do not add a separate legend.
 - **scorecard** — `factors` (row labels, including group headers), `is_header` boolean
   flags per row, `companies` (column headers), and `values` as a 3D array: outer = rows,
   middle = columns, inner = `[measure, score]` or `[]` for header rows.
+  > **SCORECARD CONTRACT — READ CAREFULLY:**
+  > - `companies` must list **the target company first, followed by peer entities** (e.g.
+  >   `["Boeing", "Airbus", "RTX", "Lockheed Martin"]`). Never put two time-horizons of the
+  >   same company here — that produces a scorecard with no peers. The first entry is the
+  >   target; its LTM scorecard data goes at `values[row][1]`.
+  > - `values[row]` is **1-indexed against `companies`**: index `0` in every row is always `[]`
+  >   (a silent placeholder the builder skips). `companies[0]` maps to `values[row][1]`,
+  >   `companies[1]` maps to `values[row][2]`, and so on. Omitting the `[]` at index 0 will
+  >   shift every peer column one position and silently misalign the data.
+  > - Header rows (`is_header=true`) use `values[row] = [[], [], [], ...]` — one `[]` per company
+  >   plus one for the placeholder. Length must equal `len(companies) + 1`.
+  > - **Quick checklist before writing the scorecard payload:**
+  >   1. `len(companies)` = number of peer entities (not counting the target).
+  >   2. Every non-header `values[row]` has length `len(companies) + 1`.
+  >   3. `values[row][0]` is always `[]`.
+  >   4. `values[row][i+1]` contains `["metric_value", "ScoreLabel"]` for `companies[i]`.
 - **esg_analysis** — table of CIS/E/S/G scores plus 3-5 commentary bullets.
 
 Target first in every peer table.
 
 ---
 
-## Step 4 — Build the editable .pptx
+## Step 4 — Build the HTML report
 
 **Default output location: always save runs to the user's Desktop** so they're easy to
 find. Use `~/Desktop/rating-pitch/<company>-<YYYYMMDD-HHMMSS>/` as the `<output-dir>`.
 Only use a different path if the user explicitly asks for one.
 
 1. Save your resolved payload to `<output-dir>/payload.json`.
-2. Ensure `python-pptx` is available. If the user doesn't have it:
+2. No additional Python packages are required — `build_html.py` uses only the standard
+   library. Verify Python 3 is available:
    ```bash
-   python3 -m venv .venv && .venv/bin/pip install -r <skill-dir>/scripts/requirements.txt
+   python3 --version
    ```
-   or simply `pip install python-pptx` if their environment allows it.
 3. Run the builder:
    ```bash
-   python3 <skill-dir>/scripts/build_pptx.py <output-dir>/payload.json <output-dir>/rating_pitch.pptx
+   python3 <skill-dir>/scripts/build_html.py <output-dir>/payload.json <output-dir>/rating_pitch.html
    ```
-4. Open the deck: `open <output-dir>/rating_pitch.pptx`
-5. The final assistant message gives the full `<output-dir>/rating_pitch.pptx` path so the
-   user can open the editable deck.
+4. Open the report: `open <output-dir>/rating_pitch.html`
+5. The final assistant message gives the full `<output-dir>/rating_pitch.html` path so the
+   user can open the report in their browser.
 
 If any section data is missing, still include the section in the payload (empty arrays
 are fine) — the builder handles empties gracefully and the deck will stay well-formed.
@@ -225,7 +246,7 @@ are fine) — the builder handles empties gracefully and the deck will stay well
   "currency": "USD",
   "companies": ["Boeing", "RTX", "Northrop Grumman", "..."],
   "sources": [
-    {"title": "", "source": "", "date": "", "url": "", "tool": ""}
+    {"id": 1, "title": "", "source": "", "date": "", "url": ""}  // id optional; rendered as [n] citation
   ],
   "sections": {
     "sector_overview": {
@@ -345,10 +366,10 @@ are fine) — the builder handles empties gracefully and the deck will stay well
         "..."
       ],
       "is_header": [true, false, true, false],
-      "companies": ["Boeing", "Airbus", "Lockheed"],
+      "companies": ["<TARGET>", "<PEER_1>", "<PEER_2>"],
       "values": [
-        [[], [], []],
-        [[], ["73.3", "Aaa"], ["69.2", "Aaa"], ["71.0", "Aaa"]]
+        [[], [], [], []],
+        [[], ["<target_rev>", "<target_score>"], ["<peer1_rev>", "<peer1_score>"], ["<peer2_rev>", "<peer2_score>"]]
       ]
     },
     "esg_analysis": {
@@ -364,45 +385,59 @@ are fine) — the builder handles empties gracefully and the deck will stay well
 
 ---
 
-## Deck structure
+## Report structure
 
-The Python builder emits these 23 slides, in this order:
+The Python builder emits these 26 sections as HTML, in this order:
 
 1. Cover
-2. Part 1 divider
-3. Sector Overview (3-column chips)
-4. Moody's View (outlook text + positioning + outlook pie)
-5. Global Macro Outlook (GDP table + takeaways)
-6. Rating Actions YTD (table)
-7. Part 2 divider
-8. Financial Analysis (bullets + rating history + rating trajectory line)
-9. Revenue Distribution (two pies + commentary)
-10. SWOT (2×2)
-11. Key Financial Metrics (four bar charts)
-12. Strategic Updates (2 columns)
-13. News, M&A & Leadership
-14. External Trends, Pressures & Risks
-15. Part 3 divider
-16. Peer Comparison Summary (table + commentary)
-17. Detailed Peer Comparison (wide financial table)
-18. Peer Comparison — Debt (two horizontal bar charts)
-19. Peer Comparison — Profitability (two horizontal bar charts)
-20. Peer Scatter Plots (two scatters)
-21. Scorecard Comparison (multi-column factor table)
-22. ESG Analysis (table + commentary)
-23. Sources (appendix)
-
-The builder's brand palette: `NAVY=#0A1264`, `BRIGHT=#005EFF`, `MID=#35397E`,
-`PURPLE=#7677A7`, `PALE=#A2A2C4`, `VERY_PALE=#CFD0E1`, `PINK=#D9017A`, `TEAL=#4FB3AE`,
-`GOLD=#C4A51A`, `LIGHT_GRAY=#D7D8D7`. Outlook pie palette (case-insensitive):
-`Stable → LIGHT_GRAY`, `Positive → TEAL`, `Negative → GOLD`, `Under Review → BRIGHT`.
+2. Agenda
+3. Part 1 divider
+4. Sector Overview (3-column chips)
+5. Moody's View (outlook text + positioning + outlook pie)
+6. Global Macro Outlook (GDP table + takeaways)
+7. Rating Actions YTD (table)
+8. Part 2 divider
+9. Financial Analysis (bullets + rating history line chart + rating rationale)
+10. Revenue Distribution (two pie charts + commentary)
+11. SWOT (2×2)
+12. Key Financial Metrics (four bar charts in 2×2 grid)
+13. Strategic Updates (2 columns)
+14. News, M&A & Leadership
+15. External Trends, Pressures & Risks
+16. Part 3 divider
+17. Peer Comparison Summary (table + commentary)
+18. Detailed Peer Comparison (wide financial table)
+19. Peer Comparison — Debt (two horizontal bar charts)
+20. Peer Comparison — Profitability (two horizontal bar charts)
+21. Peer Scatter Plots (two scatter charts)
+22. Scorecard Comparison (multi-column factor table)
+23. ESG Analysis (table + commentary)
+24. Citations (appendix — canonical numbered [n] references with hyperlinked titles)
+25. Thank You
+26. Disclaimer
+
+The builder's data-visualization palette (Moody's official, priority order):
+`#1 BRIGHT_BLUE=#005eff`, `#2 TEAL=#5eb6bc`, `#3 GOLD=#c7ab21`, `#4 MID_BLUE=#5c068c`,
+`#5 PINK=#ba0168`, `#6 PURPLE=#c64809`, `#7 PALE=#bed6ff`, `NAVY=#040826`,
+`LIGHT_GRAY=#e1e2e1`.
+
+Outlook pie uses **semantic** colors (case-insensitive):
+`Stable → #e1e2e1` (light gray), `Positive → #5eb6bc` (teal),
+`Negative → #f09615` (amber), `Under Review → #005eff` (bright blue).
+
+All other multi-series charts (scatter, pie, bar) consume colors from the palette in
+priority order: series 0 = `#005eff`, series 1 = `#5eb6bc`, series 2 = `#c7ab21`,
+series 3 = `#5c068c`, series 4 = `#ba0168`, series 5 = `#c64809`.
+
+Charts are rendered client-side via Chart.js (loaded from cdnjs.cloudflare.com CDN).
+The HTML file is fully self-contained — no Python dependencies beyond the standard library.
 
 ---
 
 ## Tips
 
 - Run ALL data-gathering tool calls in a single parallel batch.
-- Keep the target company first in every peer table — the `.pptx` and the
+- Keep the target company first in every peer table — the HTML report and the
   commentary all assume this ordering.
 - `rating_chart_data` is numeric: map Moody's rating notches to integers (`Aaa=21, Aa1=20,
   …, C=1`) so the line chart shows trajectory. Both `rating_history` and
@@ -411,15 +446,22 @@ The builder's brand palette: `NAVY=#0A1264`, `BRIGHT=#005EFF`, `MID=#35397E`,
 - Pie percentages must sum to 100 — bucket small categories into "Other".
 - `key_metrics` arrays must match `periods` length. Use `null` for missing points.
 - Scorecard header rows use `is_header=true` and `values[row] = [[], [], ...]` (empty
-  per-company entries). The builder turns these into merged header rows in the `.pptx`.
+  per-company entries). The builder turns these into highlighted header rows in the HTML table.
+- **Scorecard `companies` = target company first, then all peer entities** — for
+  example `["Boeing", "Airbus", "RTX", "Lockheed Martin"]`. Never put two time-horizons
+  of the same company here. `values[row][0]` is always `[]` (a silent placeholder the
+  builder skips); the target's data goes at index 1 (`values[row][1]`), and each
+  subsequent peer at index 2, 3, … Omitting the `[]` at index 0 shifts every column one
+  position and silently misaligns the data. Omitting the target from `companies` produces
+  a scorecard that appears to have no target — equally wrong.
 - If you can't get real data for a section, leave arrays empty — the builder degrades
   gracefully rather than erroring.
-- Dates in the deck are just strings; format however reads best (e.g., "Nov 20, 2025").
+- Dates in the report are just strings; format however reads best (e.g., "Nov 20, 2025").
 - The `<output-dir>` name should be lower-cased and hyphen-joined (e.g.
   `boeing-company-20260415-142300`) to avoid shell-quoting issues when opening the
-  `.pptx`.
-- Revenue value labels must use comma-separated thousands with zero decimal places
-  (`#,##0`). Use `"#,##0"` as the `y_format` argument in `add_bar_chart` for the `.pptx`.
+  `.html` file.
+- Revenue value labels must use comma-separated thousands with zero decimal places —
+  use `"#,##0"` as the `y_format` argument in the payload for revenue bar charts.
 - **Never copy `period` values from `sample_payload.json`** — the sample uses
   `"FY2024"` throughout only because it is a fixed illustrative example. In a real
   run, read the period label from the `getEntityFinancials` response for each
diff --git a/plugins/moody-s/skills/moody-s-rating-analysis/agents/openai.yaml b/plugins/moody-s/skills/moody-s-rating-analysis/agents/openai.yaml
index 8e940124..274d7159 100644
--- a/plugins/moody-s/skills/moody-s-rating-analysis/agents/openai.yaml
+++ b/plugins/moody-s/skills/moody-s-rating-analysis/agents/openai.yaml
@@ -1,4 +1,4 @@
 interface:
-  display_name: "Moody's Rating Pitch Deck"
-  short_description: "Generate an editable PowerPoint rating pitch deck covering sector analysis, company financials, SWOT, peer comparison, and ESG"
-  default_prompt: "Use Moody's Rating Pitch skill to resolve the target company, gather all credit and sector data, and build an editable Rating Pitch Deck (.pptx)."
+  display_name: "Moody's Rating Pitch"
+  short_description: "Generate a self-contained HTML rating pitch report covering sector analysis, company financials, SWOT, peer comparison, and ESG"
+  default_prompt: "Use Moody's Rating Pitch skill to resolve the target company, gather all credit and sector data, and build a Rating Pitch HTML report."
diff --git a/plugins/moody-s/skills/moody-s-rating-analysis/evals/evals.json b/plugins/moody-s/skills/moody-s-rating-analysis/evals/evals.json
index 16777476..04d9247f 100644
--- a/plugins/moody-s/skills/moody-s-rating-analysis/evals/evals.json
+++ b/plugins/moody-s/skills/moody-s-rating-analysis/evals/evals.json
@@ -3,63 +3,80 @@
   "evals": [
     {
       "id": 1,
-      "prompt": "Create a rating pitch deck for Boeing Company in the Aerospace/Defense sector",
-      "expected_output": "An editable rating_pitch.pptx (>=500KB, ~20+ slides) built from the Moody's 2026 Corp template via scripts/build_pptx.py, plus the resolved payload.json saved alongside it under ~/Desktop/rating-pitch/<company>-<YYYYMMDD-HHMMSS>/. The deck covers cover, 3-part structure (Sector / Company / Peers), macro GDP table, rating actions, SWOT (4 quadrants), key metrics charts, peer summary/financials/debt/profitability/scatter tables, scorecard, ESG, and sources.",
+      "prompt": "Create a rating pitch report for Boeing Company in the Aerospace/Defense sector",
+      "expected_output": "Two output files: (a) rating_pitch.html — a single self-contained HTML report built from scripts/build_html.py containing all 26 sections (cover, agenda, 3 part dividers, 21 content sections, sources, thank-you, disclaimer) with inline Chart.js charts, styled tables, and bullet lists using the Moody's brand palette; (b) payload.json — the fully resolved data payload. No PPTX is produced. The HTML file opens directly in a browser with no additional dependencies beyond an internet connection for the Chart.js CDN.",
       "files": [],
       "assertions": [
-        "Both artifacts exist (rating_pitch.pptx, payload.json)",
-        "PowerPoint file built from Moody's corporate template (>= 500 KB)",
-        "PowerPoint has >= 20 slides",
+        "Both output files exist: rating_pitch.html and payload.json",
+        "rating_pitch.html is a valid standalone HTML document (starts with <!DOCTYPE html> and ends with </html>)",
+        "rating_pitch.html contains >= 20 <section> elements",
+        "rating_pitch.html contains all 3 part divider slides (Sector Analysis, Company Credit Overview, Company Positioning vs. Peers)",
+        "rating_pitch.html contains >= 14 <canvas> elements (one per Chart.js chart)",
+        "rating_pitch.html contains >= 14 new Chart( calls in <script> blocks",
+        "rating_pitch.html contains >= 6 <table> elements covering GDP, rating actions, peer summary, peer financials, scorecard, ESG",
+        "SWOT section present with 4 quadrants (Strengths, Weaknesses, Opportunities, Threats) each containing >= 2 <li> bullets",
+        "Peer summary table has >= 3 company columns",
+        "Peer financials table has >= 3 company columns and >= 5 financial metric rows",
+        "Macro GDP table has >= 3 country rows",
+        "Rating history line chart present with >= 2 data points",
+        "Scorecard table has >= 3 factor rows",
+        "ESG table has >= 1 data row",
+        "Sources section present with >= 5 listed references",
         "payload.json matches the skill schema (report_date, target_company, sector, currency, companies, sources, sections)",
-        "payload.sections.swot has >= 2 items in each of strengths/weaknesses/opportunities/threats",
-        "payload.sections.peer_summary.table has >= 3 rows",
-        "payload.sections.peer_financials.rows has >= 3 rows",
-        "payload.sections.macro_outlook.gdp_table.rows has >= 3 rows",
-        "payload.sections.financial_analysis.rating_history has >= 2 rows",
-        "payload.sections.scorecard.factors has >= 3 entries",
-        "payload.sections.esg_analysis.table has >= 1 row",
-        "payload.sources contains >= 5 references"
+        "Peer debt and profitability bar charts include period labels on each bar (e.g. FY2024) and a value-axis title with unit"
       ]
     },
     {
       "id": 2,
-      "prompt": "I need a rating pitch for Tesla Inc. \u2014 sector is Automotive. Use USD and compare against 6 peers",
-      "expected_output": "Same structural requirements as eval-1, with 6 peers (not 5) resolved via getEntityPeers, USD honored in payload.currency, and peer tables / scatters reflecting 6 comparators. Tests natural-language phrasing + parameter handling (currency, peer count).",
+      "prompt": "I need a rating pitch for Tesla Inc. — sector is Automotive. Use USD and compare against 6 peers",
+      "expected_output": "Same structural requirements as eval-1, with 6 peers resolved via getEntityPeers, USD honoured in payload.currency, and peer tables/charts reflecting 6 comparators. Tests natural-language phrasing and parameter handling (currency, peer count). Output is rating_pitch.html + payload.json only — no PPTX.",
       "files": [],
       "assertions": [
-        "Both artifacts exist (rating_pitch.pptx, payload.json)",
-        "PowerPoint file built from Moody's corporate template (>= 500 KB)",
-        "PowerPoint has >= 20 slides",
-        "payload.currency == 'USD'",
-        "payload.companies has >= 6 entries (target + 6 peers, target first)",
-        "payload.sections.peer_summary.table has >= 6 rows",
-        "payload.sections.peer_financials.rows has >= 6 rows",
-        "payload.sections.swot has >= 2 items in each of strengths/weaknesses/opportunities/threats",
-        "payload.sections.macro_outlook.gdp_table.rows has >= 3 rows",
-        "payload.sections.financial_analysis.rating_history has >= 2 rows",
-        "payload.sections.scorecard.factors has >= 3 entries",
-        "payload.sections.esg_analysis.table has >= 1 row",
-        "payload.sources contains >= 5 references"
+        "Both output files exist: rating_pitch.html and payload.json",
+        "rating_pitch.html is a valid standalone HTML document (starts with <!DOCTYPE html> and ends with </html>)",
+        "rating_pitch.html contains >= 20 <section> elements",
+        "rating_pitch.html contains all 3 part divider slides (Sector Analysis, Company Credit Overview, Company Positioning vs. Peers)",
+        "rating_pitch.html contains >= 14 <canvas> elements (one per Chart.js chart)",
+        "rating_pitch.html contains >= 14 new Chart( calls in <script> blocks",
+        "rating_pitch.html contains >= 6 <table> elements",
+        "SWOT section present with 4 quadrants (Strengths, Weaknesses, Opportunities, Threats) each containing >= 2 <li> bullets",
+        "Peer summary table has >= 6 company columns (6 peers requested)",
+        "Peer financials table has >= 6 company columns and >= 5 financial metric rows",
+        "Macro GDP table has >= 3 country rows",
+        "Rating history line chart present with >= 2 data points",
+        "Scorecard table has >= 3 factor rows",
+        "ESG table has >= 1 data row",
+        "Sources section present with >= 5 listed references",
+        "payload.json currency field equals USD",
+        "payload.json companies array contains >= 6 peer entries",
+        "payload.json matches the skill schema (report_date, target_company, sector, currency, companies, sources, sections)",
+        "Peer debt and profitability bar charts include period labels on each bar and a value-axis title with unit"
       ]
     },
     {
       "id": 3,
-      "prompt": "Generate a credit pitch deck for JPMorgan Chase, Banking sector",
-      "expected_output": "Same structural requirements as eval-1, applied to a financial-institution target where bank-specific credit metrics (deposits, capital ratios) substitute for generic Debt/EBITDA. Tests sector flex (Banking vs A&D/Auto).",
+      "prompt": "Generate a credit pitch report for JPMorgan Chase, Banking sector",
+      "expected_output": "Same structural requirements as eval-1, applied to a financial-institution target where bank-specific credit metrics (deposits, capital ratios, NIM) may substitute for generic Debt/EBITDA where appropriate. Tests sector flexibility (Banking vs A&D/Auto). Output is rating_pitch.html + payload.json only — no PPTX.",
       "files": [],
       "assertions": [
-        "Both artifacts exist (rating_pitch.pptx, payload.json)",
-        "PowerPoint file built from Moody's corporate template (>= 500 KB)",
-        "PowerPoint has >= 20 slides",
-        "payload.sector matches 'Banking' (case-insensitive)",
-        "payload.sections.swot has >= 2 items in each of strengths/weaknesses/opportunities/threats",
-        "payload.sections.peer_summary.table has >= 3 rows",
-        "payload.sections.peer_financials.rows has >= 3 rows",
-        "payload.sections.macro_outlook.gdp_table.rows has >= 3 rows",
-        "payload.sections.financial_analysis.rating_history has >= 2 rows",
-        "payload.sections.scorecard.factors has >= 3 entries",
-        "payload.sections.esg_analysis.table has >= 1 row",
-        "payload.sources contains >= 5 references"
+        "Both output files exist: rating_pitch.html and payload.json",
+        "rating_pitch.html is a valid standalone HTML document (starts with <!DOCTYPE html> and ends with </html>)",
+        "rating_pitch.html contains >= 20 <section> elements",
+        "rating_pitch.html contains all 3 part divider slides (Sector Analysis, Company Credit Overview, Company Positioning vs. Peers)",
+        "rating_pitch.html contains >= 14 <canvas> elements (one per Chart.js chart)",
+        "rating_pitch.html contains >= 14 new Chart( calls in <script> blocks",
+        "rating_pitch.html contains >= 6 <table> elements",
+        "SWOT section present with 4 quadrants (Strengths, Weaknesses, Opportunities, Threats) each containing >= 2 <li> bullets",
+        "Peer summary table has >= 3 company columns",
+        "Peer financials table has >= 3 company columns and >= 5 financial metric rows",
+        "Macro GDP table has >= 3 country rows",
+        "Rating history line chart present with >= 2 data points",
+        "Scorecard table has >= 3 factor rows",
+        "ESG table has >= 1 data row",
+        "Sources section present with >= 5 listed references",
+        "payload.json target_company field contains JPMorgan and sector field contains Banking",
+        "payload.json matches the skill schema (report_date, target_company, sector, currency, companies, sources, sections)",
+        "Peer debt and profitability bar charts include period labels on each bar and a value-axis title with unit"
       ]
     }
   ]
diff --git a/plugins/moody-s/skills/moody-s-rating-analysis/scripts/build_pptx.py b/plugins/moody-s/skills/moody-s-rating-analysis/scripts/build_pptx.py
deleted file mode 100644
index 51ded150..00000000
--- a/plugins/moody-s/skills/moody-s-rating-analysis/scripts/build_pptx.py
+++ /dev/null
@@ -1,1690 +0,0 @@
-#!/usr/bin/env python3
-"""Build a Moody's Rating Analysis Deck by populating the official Moody's Corp
-Template PPTX (assets/Moody_Corp_Template.pptx).
-
-Rather than building slides from scratch, this script opens the branded
-template, removes its demo slides, and adds slides that use the template's
-built-in layouts (Cover, Agenda, Divider, 1/2/3 Column, Back Cover,
-Disclaimer). All fonts, colors, and decorative elements come from the
-template's master slides — we only fill placeholders and add charts/tables.
-
-Usage:
-    python build_pptx.py <payload.json> <output.pptx>
-"""
-from __future__ import annotations
-
-import json
-import sys
-from copy import deepcopy
-from pathlib import Path
-
-from pptx import Presentation
-from pptx.chart.data import CategoryChartData, XyChartData
-from pptx.dml.color import RGBColor
-from pptx.enum.chart import (XL_CHART_TYPE, XL_LABEL_POSITION,
-                             XL_LEGEND_POSITION, XL_TICK_LABEL_POSITION)
-from pptx.enum.text import PP_ALIGN, MSO_ANCHOR, MSO_AUTO_SIZE
-from pptx.oxml.ns import qn
-from pptx.util import Emu, Inches, Pt
-
-
-TEMPLATE_PATH = (Path(__file__).resolve().parent.parent
-                 / "assets" / "Moody_Corp_Template.pptx")
-
-
-# ---- Layout indices (from Moody's Corp Template_2026) ----------------------
-
-L_COVER = 0            # Cover 1
-L_AGENDA = 10          # Agenda 4 (3 object columns)
-L_DIVIDER = 11         # Divider 1 - Short title
-L_ONE_COL = 24         # 1 Column - Subhead
-L_TWO_COL = 25         # 2 Column, Equal - Subhead
-L_THREE_COL = 27       # 3 Column - Subhead
-L_FOUR_COL = 28        # 4 Column - Subhead
-L_TITLE_ONLY = 44      # Title Only
-L_BACK_COVER = 57      # Back Cover 1 (Thank you)
-L_DISCLAIMER = 63      # Disclaimer
-
-# Brand palette used for chart accents (matches template theme colors).
-NAVY = RGBColor(0x0A, 0x12, 0x64)
-BRIGHT_BLUE = RGBColor(0x00, 0x5E, 0xFF)
-MID_BLUE = RGBColor(0x35, 0x39, 0x7E)
-PURPLE_BLUE = RGBColor(0x76, 0x77, 0xA7)
-PALE_PURPLE = RGBColor(0xA2, 0xA2, 0xC4)
-VERY_PALE = RGBColor(0xCF, 0xD0, 0xE1)
-PINK = RGBColor(0xD9, 0x01, 0x7A)
-LIGHT_GRAY = RGBColor(0xD7, 0xD8, 0xD7)
-TEAL = RGBColor(0x4F, 0xB3, 0xAE)
-GOLD = RGBColor(0xC4, 0xA5, 0x1A)
-GRID_LINE = RGBColor(0xCF, 0xD0, 0xE1)
-
-PIE_PALETTE = [BRIGHT_BLUE, TEAL, GOLD, MID_BLUE, PURPLE_BLUE, PINK,
-               PALE_PURPLE, VERY_PALE]
-OUTLOOK_PALETTE = {
-    "stable": LIGHT_GRAY,
-    "positive": TEAL,
-    "negative": GOLD,
-    "under review": BRIGHT_BLUE,
-}
-
-TEXT_DARK = RGBColor(0x22, 0x22, 0x22)
-TEXT_MUTED = RGBColor(0x66, 0x6B, 0x73)
-WHITE = RGBColor(0xFF, 0xFF, 0xFF)
-HEADER_FILL = RGBColor(0xE9, 0xEA, 0xEE)
-
-# Full Moody's rating scale: numeric notch → label (Aaa=21 … C=1)
-RATING_SCALE = {
-    21: "Aaa",
-    20: "Aa1", 19: "Aa2", 18: "Aa3",
-    17: "A1",  16: "A2",  15: "A3",
-    14: "Baa1", 13: "Baa2", 12: "Baa3",
-    11: "Ba1",  10: "Ba2",   9: "Ba3",
-     8: "B1",    7: "B2",    6: "B3",
-     5: "Caa1",  4: "Caa2",  3: "Caa3",
-     2: "Ca",    1: "C",
-}
-
-def notch_to_label(n):
-    return RATING_SCALE.get(int(round(n)), str(int(round(n))))
-
-
-# ---- Utility helpers -------------------------------------------------------
-
-def remove_all_slides(prs):
-    """Keep the template's masters/layouts; drop all existing slides."""
-    id_lst = prs.slides._sldIdLst
-    for sld in list(id_lst):
-        rid = sld.attrib[qn("r:id")]
-        prs.part.drop_rel(rid)
-        id_lst.remove(sld)
-
-
-_TAG_RE = __import__("re").compile(r"<[^>]+>")
-
-
-def html_to_text(s):
-    """Strip HTML tags and unescape a small set of entities. Paragraphs and
-    list items are converted to blank-line separated / bulleted plain text."""
-    if s is None:
-        return ""
-    s = str(s)
-    s = (s.replace("</p>", "\n\n")
-          .replace("<br>", "\n")
-          .replace("<br/>", "\n")
-          .replace("<br />", "\n")
-          .replace("<li>", "• ")
-          .replace("</li>", "\n"))
-    s = _TAG_RE.sub("", s)
-    s = (s.replace("&amp;", "&")
-          .replace("&nbsp;", " ")
-          .replace("&rarr;", "→")
-          .replace("&#39;", "'")
-          .replace("&quot;", '"')
-          .replace("&lt;", "<")
-          .replace("&gt;", ">"))
-    return "\n".join(line.rstrip() for line in s.splitlines()).strip()
-
-
-def get_ph(slide, idx):
-    for ph in slide.placeholders:
-        if ph.placeholder_format.idx == idx:
-            return ph
-    return None
-
-
-def del_ph(slide, idx):
-    ph = get_ph(slide, idx)
-    if ph is not None:
-        ph._element.getparent().remove(ph._element)
-
-
-def clear_empty_placeholders(slide):
-    """Remove any placeholder that was left empty so PowerPoint doesn't show
-    'Click to add text' prompts."""
-    for ph in list(slide.placeholders):
-        try:
-            if ph.has_text_frame and not ph.text_frame.text.strip():
-                ph._element.getparent().remove(ph._element)
-        except Exception:
-            pass
-
-
-def ph_bounds(slide, idx):
-    ph = get_ph(slide, idx)
-    if ph is None:
-        return None
-    return (ph.left, ph.top, ph.width, ph.height)
-
-
-def set_text(slide, idx, text):
-    ph = get_ph(slide, idx)
-    if ph is None:
-        return None
-    cleaned = html_to_text(text)
-    if not cleaned:
-        del_ph(slide, idx)
-        return None
-    tf = ph.text_frame
-    tf.text = cleaned
-    if idx == 0:
-        try:
-            tf.word_wrap = True
-            tf.auto_size = MSO_AUTO_SIZE.TEXT_TO_FIT_SHAPE
-            size_pt = None
-            if len(cleaned) > 60:
-                size_pt = 18
-            elif len(cleaned) > 45:
-                size_pt = 22
-            elif len(cleaned) > 35:
-                size_pt = 26
-            if size_pt is not None:
-                for p in tf.paragraphs:
-                    for r in p.runs:
-                        r.font.size = Pt(size_pt)
-        except Exception:
-            pass
-    return ph
-
-
-def title_bottom(slide, default=Inches(1.1), text=None):
-    """Return the bottom edge (EMU) of the title placeholder so downstream
-    content can be positioned beneath it regardless of template layout.
-    If ``text`` is provided, adds a safety buffer when the title is likely
-    to wrap based on a conservative chars-per-line estimate."""
-    ph = get_ph(slide, 0)
-    if ph is None:
-        return default
-    try:
-        bot = ph.top + ph.height
-        if text:
-            chars_per_line = max(20, int(ph.width / Inches(0.17)))
-            extra_lines = max(0, (len(text) - 1) // chars_per_line)
-            if extra_lines > 0:
-                bot += Inches(0.45) * extra_lines
-        return bot
-    except Exception:
-        return default
-
-
-def set_bullets(slide, idx, items, *, bullet_char=None, size=None):
-    """Populate a placeholder with multiple paragraphs. Formatting comes
-    from the layout/master — we only set the text content."""
-    ph = get_ph(slide, idx)
-    if ph is None:
-        return None
-    tf = ph.text_frame
-    tf.word_wrap = True
-    cleaned = [html_to_text(x) for x in (items or []) if x is not None]
-    cleaned = [c for c in cleaned if c]
-    if not cleaned:
-        del_ph(slide, idx)
-        return None
-    tf.text = cleaned[0]
-    for item in cleaned[1:]:
-        p = tf.add_paragraph()
-        p.text = item
-    if size is not None:
-        for p in tf.paragraphs:
-            for r in p.runs:
-                r.font.size = Pt(size)
-    return ph
-
-
-def set_bold_prefix_bullets(slide, idx, items, *, size=None):
-    """Items are [{bold, text}, ...] — each becomes one paragraph with the
-    prefix in bold followed by the rest of the text."""
-    ph = get_ph(slide, idx)
-    if ph is None:
-        return None
-    tf = ph.text_frame
-    tf.word_wrap = True
-    tf.text = ""
-    p0 = tf.paragraphs[0]
-    for i, item in enumerate(items or []):
-        if isinstance(item, dict):
-            prefix = (item.get("bold") or item.get("prefix")
-                      or item.get("date") or "")
-            rest = (item.get("text") or item.get("summary") or "")
-        elif isinstance(item, (tuple, list)) and len(item) >= 2:
-            prefix, rest = str(item[0]), str(item[1])
-        else:
-            prefix, rest = "", str(item)
-        p = p0 if i == 0 else tf.add_paragraph()
-        # Clear any inherited run, then add prefix + rest as two runs.
-        for r in list(p.runs):
-            r._r.getparent().remove(r._r)
-        if prefix:
-            r1 = p.add_run()
-            r1.text = prefix + (" – " if rest and not prefix.endswith(("–", ":")) else " ")
-            r1.font.bold = True
-            if size is not None:
-                r1.font.size = Pt(size)
-        if rest:
-            r2 = p.add_run()
-            r2.text = rest
-            if size is not None:
-                r2.font.size = Pt(size)
-
-
-# ---- Chart + table helpers -------------------------------------------------
-
-def _style_axes(chart, *, size=8):
-    for axis in (chart.category_axis, chart.value_axis):
-        try:
-            axis.tick_labels.font.name = "Arial"
-            axis.tick_labels.font.size = Pt(size)
-            axis.tick_labels.font.color.rgb = TEXT_MUTED
-        except Exception:
-            pass
-        try:
-            if axis.has_major_gridlines:
-                axis.major_gridlines.format.line.color.rgb = GRID_LINE
-                axis.major_gridlines.format.line.width = Pt(0.5)
-        except Exception:
-            pass
-
-
-def _style_title(chart, title, *, size=13):
-    if not title:
-        chart.has_title = False
-        return
-    chart.has_title = True
-    chart.chart_title.text_frame.text = title
-    for p in chart.chart_title.text_frame.paragraphs:
-        for r in p.runs:
-            r.font.name = "Arial"
-            r.font.size = Pt(size)
-            r.font.bold = True
-            r.font.color.rgb = TEXT_DARK
-
-
-def add_bar_chart(slide, left, top, width, height, title, categories, values,
-                  *, color=BRIGHT_BLUE, y_format=None, show_value=True):
-    cd = CategoryChartData()
-    cd.categories = [str(c) for c in categories]
-    cd.add_series(title or "Series",
-                  [float(v) if v is not None else 0.0 for v in values])
-    title_h = Inches(0.3) if title else 0
-    chart_top = top + title_h
-    chart_h = height - title_h
-    if title:
-        add_textbox(slide, left, top, width, title_h, title,
-                    size=12, bold=True, color=TEXT_DARK,
-                    align=PP_ALIGN.CENTER, anchor=MSO_ANCHOR.MIDDLE)
-    gf = slide.shapes.add_chart(XL_CHART_TYPE.COLUMN_CLUSTERED,
-                                left, chart_top, width, chart_h, cd)
-    chart = gf.chart
-    chart.has_title = False
-    chart.has_legend = False
-    plot = chart.plots[0]
-    plot.has_data_labels = show_value
-    if show_value:
-        dl = plot.data_labels
-        dl.position = XL_LABEL_POSITION.OUTSIDE_END
-        dl.font.name = "Arial"
-        dl.font.size = Pt(9)
-        dl.font.bold = True
-        dl.font.color.rgb = TEXT_DARK
-        if y_format:
-            dl.number_format = y_format
-            dl.number_format_is_linked = False
-    series = chart.series[0]
-    series.format.fill.solid()
-    series.format.fill.fore_color.rgb = color
-    series.format.line.fill.background()
-    _style_axes(chart)
-    try:
-        chart.category_axis.tick_label_position = XL_TICK_LABEL_POSITION.LOW
-    except Exception:
-        pass
-    return chart
-
-
-def add_line_chart(slide, left, top, width, height, title, categories, values,
-                   *, color=BRIGHT_BLUE, rating_scale=False):
-    cd = CategoryChartData()
-    cd.categories = [str(c) for c in categories]
-    cd.add_series(title or "Series",
-                  [float(v) if v is not None else None for v in values])
-
-    if rating_scale:
-        # Reserve left margin for custom rating label textboxes
-        label_w = Inches(0.55)
-        chart_left  = left + label_w
-        chart_width = width - label_w
-    else:
-        chart_left  = left
-        chart_width = width
-
-    title_h = Inches(0.3) if title else 0
-    if title:
-        add_textbox(slide, left, top, width, title_h, title,
-                    size=12, bold=True, color=TEXT_DARK,
-                    align=PP_ALIGN.CENTER, anchor=MSO_ANCHOR.MIDDLE)
-
-    gf = slide.shapes.add_chart(XL_CHART_TYPE.LINE,
-                                chart_left, top + title_h,
-                                chart_width, height - title_h, cd)
-    chart = gf.chart
-    chart.has_title = False
-    chart.has_legend = False
-    s = chart.series[0]
-    s.format.line.color.rgb = color
-    s.format.line.width = Pt(2)
-    try:
-        s.marker.size = 6
-        s.marker.format.fill.solid()
-        s.marker.format.fill.fore_color.rgb = color
-        s.marker.format.line.color.rgb = color
-    except Exception:
-        pass
-    _style_axes(chart)
-
-    if rating_scale:
-        try:
-            clean = [float(v) for v in values if v is not None]
-            if clean:
-                # ── Smart scale: anchors + contextual window ──────────────
-                current = int(round(max(clean)))
-                ANCHORS = {21, 18, 15, 12, 9, 6, 3}  # Aaa,Aa3,A3,Baa3,Ba3,B3,Caa3
-                mn = max(1,  current - 5)
-                mx = min(21, current + 5)
-                context = set(range(max(1, current - 2), min(21, current + 2) + 1))
-                visible = (ANCHORS | context) & set(range(mn, mx + 1))
-
-                val_ax = chart.value_axis
-                val_ax.minimum_scale = mn
-                val_ax.maximum_scale = mx
-                val_ax.major_unit    = 1
-
-                # Hide the numeric axis labels — PowerPoint ignores conditional
-                # numFmt on value axes, so we use a textbox overlay instead.
-                try:
-                    val_ax.tick_labels.font.size = Pt(1)
-                    val_ax.tick_labels.font.color.rgb = RGBColor(0xFF, 0xFF, 0xFF)
-                except Exception:
-                    pass
-                from lxml import etree
-                ns      = "http://schemas.openxmlformats.org/drawingml/2006/chart"
-                fmt_tag = f"{{{ns}}}numFmt"
-                existing = val_ax._element.find(fmt_tag)
-                if existing is not None:
-                    existing.attrib["formatCode"]  = '""'
-                    existing.attrib["sourceLinked"] = "0"
-                else:
-                    el = etree.SubElement(val_ax._element, fmt_tag,
-                                          attrib={"formatCode": '""',
-                                                  "sourceLinked": "0"})
-                    val_ax._element.remove(el)
-                    val_ax._element.insert(1, el)
-
-                # ── Textbox overlay: place rating labels at correct Y positions ──
-                chart_top_emu   = top + title_h
-                chart_h_emu     = height - title_h
-                # Approximate PowerPoint plot area vertical bounds
-                plot_top    = chart_top_emu + int(chart_h_emu * 0.08)
-                plot_bottom = chart_top_emu + int(chart_h_emu * 0.88)
-                plot_height = plot_bottom - plot_top
-                lbl_h = Inches(0.18)
-                lbl_w = label_w - Emu(36000)
-
-                for notch in range(mn, mx + 1):
-                    if notch not in visible:
-                        continue
-                    frac     = (notch - mn) / (mx - mn)
-                    y_centre = plot_bottom - int(plot_height * frac)
-                    y        = y_centre - lbl_h // 2
-                    label    = notch_to_label(notch)
-                    is_anchor  = notch in ANCHORS
-                    is_current = notch == current
-                    tb = slide.shapes.add_textbox(left, y, lbl_w, lbl_h)
-                    tf = tb.text_frame
-                    tf.word_wrap = False
-                    tf.margin_left = tf.margin_right = 0
-                    tf.margin_top  = tf.margin_bottom = 0
-                    p = tf.paragraphs[0]
-                    p.alignment = PP_ALIGN.RIGHT
-                    r = p.add_run()
-                    r.text = label
-                    r.font.name  = "Arial"
-                    r.font.size  = Pt(9 if is_anchor else 8)
-                    r.font.bold  = is_anchor or is_current
-                    r.font.color.rgb = (NAVY if is_current
-                                        else TEXT_DARK if is_anchor
-                                        else TEXT_MUTED)
-        except Exception:
-            pass
-    return chart
-
-
-def add_pie_chart(slide, left, top, width, height, title, labels, values,
-                  *, colors=None, show_percent=True, use_legend=False,
-                  legend_position=None):
-    cd = CategoryChartData()
-    cd.categories = list(labels)
-    cd.add_series(title or "Series", [float(v or 0) for v in values])
-    title_h = Inches(0.35) if title else 0
-    if title:
-        add_textbox(slide, left, top, width, title_h, title,
-                    size=12, bold=True, color=TEXT_DARK,
-                    align=PP_ALIGN.CENTER, anchor=MSO_ANCHOR.MIDDLE)
-    gf = slide.shapes.add_chart(XL_CHART_TYPE.PIE,
-                                left, top + title_h, width,
-                                height - title_h, cd)
-    chart = gf.chart
-    chart.has_title = False
-    chart.has_legend = bool(use_legend)
-    if use_legend:
-        try:
-            chart.legend.position = legend_position or XL_LEGEND_POSITION.BOTTOM
-            chart.legend.include_in_layout = False
-            chart.legend.font.name = "Arial"
-            chart.legend.font.size = Pt(9)
-            chart.legend.font.color.rgb = TEXT_DARK
-        except Exception:
-            pass
-    plot = chart.plots[0]
-    plot.has_data_labels = True
-    dl = plot.data_labels
-    dl.show_percentage = show_percent
-    dl.show_value = not show_percent
-    dl.show_category_name = not use_legend
-    dl.font.name = "Arial"
-    dl.font.size = Pt(9)
-    dl.font.bold = True
-    dl.font.color.rgb = TEXT_DARK
-    dl.position = (XL_LABEL_POSITION.BEST_FIT if use_legend
-                   else XL_LABEL_POSITION.OUTSIDE_END)
-    palette = list(colors) if colors else PIE_PALETTE
-    for i, pt in enumerate(chart.series[0].points):
-        c = palette[i % len(palette)]
-        pt.format.fill.solid()
-        pt.format.fill.fore_color.rgb = c
-        pt.format.line.color.rgb = WHITE
-        pt.format.line.width = Pt(1)
-    return chart
-
-
-def add_scatter_chart(slide, left, top, width, height, title, points, *,
-                      x_label="", y_label=""):
-    clean = []
-    for p in points or []:
-        x = p.get("x")
-        y = p.get("y")
-        if x is None or y is None:
-            continue
-        try:
-            clean.append((p.get("company", "") or p.get("name", ""),
-                          float(x), float(y)))
-        except (TypeError, ValueError):
-            continue
-    if not clean:
-        return None, []
-    cd = XyChartData()
-    s = cd.add_series(title or "Series")
-    for _, x, y in clean:
-        s.add_data_point(x, y)
-    title_h = Inches(0.3) if title else 0
-    if title:
-        add_textbox(slide, left, top, width, title_h, title,
-                    size=12, bold=True, color=TEXT_DARK,
-                    align=PP_ALIGN.CENTER, anchor=MSO_ANCHOR.MIDDLE)
-    gf = slide.shapes.add_chart(XL_CHART_TYPE.XY_SCATTER,
-                                left, top + title_h, width,
-                                height - title_h, cd)
-    chart = gf.chart
-    chart.has_title = False
-    chart.has_legend = False
-    sr = chart.series[0]
-    sr.marker.size = 9
-    sr.marker.format.fill.solid()
-    sr.marker.format.fill.fore_color.rgb = BRIGHT_BLUE
-    sr.marker.format.line.color.rgb = BRIGHT_BLUE
-    try:
-        sr.format.line.fill.background()
-    except Exception:
-        pass
-    for ax, lab in ((chart.category_axis, x_label),
-                     (chart.value_axis, y_label)):
-        try:
-            ax.has_title = bool(lab)
-            if lab:
-                ax.axis_title.text_frame.text = lab
-                for p in ax.axis_title.text_frame.paragraphs:
-                    for r in p.runs:
-                        r.font.name = "Arial"
-                        r.font.size = Pt(9)
-                        r.font.color.rgb = TEXT_MUTED
-            ax.tick_labels.font.name = "Arial"
-            ax.tick_labels.font.size = Pt(8)
-            ax.tick_labels.font.color.rgb = TEXT_MUTED
-        except Exception:
-            pass
-        try:
-            if ax.has_major_gridlines:
-                ax.major_gridlines.format.line.color.rgb = GRID_LINE
-                ax.major_gridlines.format.line.width = Pt(0.5)
-        except Exception:
-            pass
-    return chart, clean
-
-
-def add_scatter_labels(slide, left, top, width, height, clean,
-                       *, title_offset=Inches(0.3)):
-    """Place small text labels above each scatter point. The bounds passed in
-    are the original slot (same as add_scatter_chart); we apply the same title
-    offset the chart helper uses so labels line up with the plot."""
-    if not clean:
-        return
-    chart_top = top + title_offset
-    chart_h = height - title_offset
-    xs = [c[1] for c in clean]
-    ys = [c[2] for c in clean]
-    xmin, xmax = min(xs), max(xs)
-    ymin, ymax = min(ys), max(ys)
-    if xmax == xmin:
-        xmax = xmin + 1
-    if ymax == ymin:
-        ymax = ymin + 1
-    plot_left = left + int(width * 0.14)
-    plot_top = chart_top + int(chart_h * 0.08)
-    plot_w = int(width * 0.80)
-    plot_h = int(chart_h * 0.74)
-    placed = []
-    label_w = Inches(0.95)
-    label_h = Inches(0.18)
-    for name, x, y in clean:
-        if not name:
-            continue
-        fx = (x - xmin) / (xmax - xmin)
-        fy = 1 - (y - ymin) / (ymax - ymin)
-        px = plot_left + int(plot_w * fx)
-        py = plot_top + int(plot_h * fy)
-        lx = px - label_w // 2
-        ly = py - label_h - Inches(0.08)
-        for _ in range(6):
-            collision = any(
-                abs(lx - ox) < label_w * 0.75 and abs(ly - oy) < label_h
-                for ox, oy in placed
-            )
-            if not collision:
-                break
-            ly -= label_h
-        placed.append((lx, ly))
-        tb = slide.shapes.add_textbox(lx, ly, label_w, label_h)
-        tf = tb.text_frame
-        tf.word_wrap = False
-        tf.margin_left = 0
-        tf.margin_right = 0
-        tf.margin_top = 0
-        tf.margin_bottom = 0
-        p = tf.paragraphs[0]
-        p.alignment = PP_ALIGN.CENTER
-        r = p.add_run()
-        r.text = name
-        r.font.name = "Arial"
-        r.font.size = Pt(8)
-        r.font.bold = True
-        r.font.color.rgb = TEXT_DARK
-
-
-def add_textbox(slide, left, top, width, height, text, *, size=10, bold=False,
-                color=TEXT_DARK, font="Arial", align=PP_ALIGN.LEFT,
-                anchor=MSO_ANCHOR.TOP):
-    tb = slide.shapes.add_textbox(left, top, width, height)
-    tf = tb.text_frame
-    tf.word_wrap = True
-    tf.vertical_anchor = anchor
-    p = tf.paragraphs[0]
-    p.alignment = align
-    r = p.add_run()
-    r.text = str(text or "")
-    r.font.name = font
-    r.font.size = Pt(size)
-    r.font.bold = bold
-    r.font.color.rgb = color
-    return tb
-
-
-def add_table(slide, left, top, width, height, headers, rows, *,
-              header_fill=HEADER_FILL, first_col_bold=False,
-              header_size=10, body_size=9, col_widths=None):
-    nrows = len(rows) + 1
-    ncols = len(headers)
-    shape = slide.shapes.add_table(nrows, ncols, left, top, width, height)
-    table = shape.table
-    if col_widths:
-        total = float(sum(col_widths))
-        for i, w in enumerate(col_widths):
-            table.columns[i].width = int(width * (w / total))
-    for j, h in enumerate(headers):
-        c = table.cell(0, j)
-        c.fill.solid()
-        c.fill.fore_color.rgb = header_fill
-        tf = c.text_frame
-        tf.text = str(h or "")
-        for p in tf.paragraphs:
-            for r in p.runs:
-                r.font.name = "Arial"
-                r.font.size = Pt(header_size)
-                r.font.bold = True
-                r.font.color.rgb = TEXT_DARK
-    for i, row in enumerate(rows, start=1):
-        for j, val in enumerate(row):
-            c = table.cell(i, j)
-            c.fill.solid()
-            c.fill.fore_color.rgb = WHITE
-            tf = c.text_frame
-            tf.text = str(val) if val is not None else ""
-            for p in tf.paragraphs:
-                for r in p.runs:
-                    r.font.name = "Arial"
-                    r.font.size = Pt(body_size)
-                    r.font.bold = first_col_bold and j == 0
-                    r.font.color.rgb = TEXT_DARK
-    return table
-
-
-# ---- HTML parsing for news/trends sections ---------------------------------
-
-def _parse_html_groups(html: str):
-    """Parse '<strong>Group</strong><ul><li>item</li>...</ul>' chunks."""
-    if not html:
-        return []
-    out = []
-    for chunk in html.split("<strong")[1:]:
-        _, _, rest = chunk.partition(">")
-        name, _, tail = rest.partition("</strong>")
-        items = []
-        if "<ul>" in tail and "</ul>" in tail:
-            ul = tail.split("<ul>", 1)[1].split("</ul>", 1)[0]
-            for li in ul.split("<li>")[1:]:
-                t = li.split("</li>", 1)[0].strip()
-                t = (t.replace("&amp;", "&")
-                      .replace("&nbsp;", " ")
-                      .replace("&rarr;", "→")
-                      .replace("&#39;", "'"))
-                if t:
-                    items.append(t)
-        name = name.replace("&amp;", "&").strip()
-        out.append({"name": name, "items": items})
-    return out
-
-
-def _normalize_categorized(section):
-    """Return list of {name, items} from string/list/dict input."""
-    if isinstance(section, str):
-        return _parse_html_groups(section)
-    if isinstance(section, list):
-        out = []
-        for g in section:
-            if isinstance(g, dict):
-                out.append({
-                    "name": g.get("name") or g.get("category") or "",
-                    "items": g.get("items") or [],
-                })
-        return out
-    if isinstance(section, dict):
-        if section.get("groups"):
-            return _normalize_categorized(section["groups"])
-        if section.get("html"):
-            return _parse_html_groups(section["html"])
-    return []
-
-
-# ---- Slide builders --------------------------------------------------------
-
-def build_cover(prs, data):
-    s = prs.slides.add_slide(prs.slide_layouts[L_COVER])
-    company = data.get("target_company", "")
-    set_text(s, 0, company)
-    set_text(s, 100, "Rating Analysis DECK")
-    set_text(s, 101, data.get("sector", ""))
-    set_text(s, 102, f"Report date: {data.get('report_date', '')}")
-    set_text(s, 99, "MOODY'S ANALYTICS")
-    return s
-
-
-def build_agenda(prs, data):
-    s = prs.slides.add_slide(prs.slide_layouts[L_AGENDA])
-    agenda = [
-        ("01", "Sector Analysis",
-         "Macro context, sector outlook and recent rating actions"),
-        ("02", "Company Credit Overview",
-         "Financial profile, SWOT, strategy and news"),
-        ("03", "Company Positioning vs. Peers",
-         "Benchmarking on leverage, profitability, scorecard and ESG"),
-    ]
-    for i, (num, title, desc) in enumerate(agenda):
-        ph = get_ph(s, 18 + i)
-        if ph is None:
-            continue
-        tf = ph.text_frame
-        tf.word_wrap = True
-        tf.text = ""
-        p0 = tf.paragraphs[0]
-        r = p0.add_run()
-        r.text = num
-        r.font.name = "Times New Roman"
-        r.font.size = Pt(64)
-        r.font.bold = True
-        r.font.color.rgb = BRIGHT_BLUE
-        p1 = tf.add_paragraph()
-        r1 = p1.add_run()
-        r1.text = title
-        r1.font.name = "Arial"
-        r1.font.size = Pt(16)
-        r1.font.bold = True
-        r1.font.color.rgb = TEXT_DARK
-        p2 = tf.add_paragraph()
-        r2 = p2.add_run()
-        r2.text = desc
-        r2.font.name = "Arial"
-        r2.font.size = Pt(10)
-        r2.font.color.rgb = TEXT_MUTED
-    return s
-
-
-def build_divider(prs, title):
-    s = prs.slides.add_slide(prs.slide_layouts[L_DIVIDER])
-    set_text(s, 0, title.upper())
-    return s
-
-
-def build_sector_overview(prs, data):
-    """3-column layout: Overview | Watchlist | Takeaways."""
-    s = prs.slides.add_slide(prs.slide_layouts[L_THREE_COL])
-    sec = data.get("sections", {}).get("sector_overview", {}) or {}
-    sector = data.get("sector", "")
-    set_text(s, 0, f"{sector} – Sector overview")
-    set_text(s, 21, "")
-    set_text(s, 29, "SECTOR OVERVIEW")
-    set_text(s, 30, "WHAT TO LOOK OUT FOR?")
-    set_text(s, 31, "KEY TAKEAWAYS")
-    set_text(s, 19, "Source: Moodys.com")
-    # Convert object placeholders into text content.
-    set_bullets(s, 23, sec.get("overview_bullets", []), size=11)
-    set_bullets(s, 24, sec.get("watchlist_bullets", []), size=11)
-    set_bullets(s, 25, sec.get("takeaway_bullets", []), size=11)
-    return s
-
-
-def build_moodys_view(prs, data):
-    s = prs.slides.add_slide(prs.slide_layouts[L_TWO_COL])
-    mv = data.get("sections", {}).get("moodys_view", {}) or {}
-    sector = data.get("sector", "")
-    title_text = f"{sector} – Moody's view"
-    set_text(s, 0, title_text)
-    set_text(s, 19, "Source: Moodys.com")
-    del_ph(s, 21)
-    del_ph(s, 29)
-    del_ph(s, 30)
-    bnd_l = ph_bounds(s, 27)
-    bnd_r = ph_bounds(s, 28)
-    del_ph(s, 27)
-    del_ph(s, 28)
-
-    summary = html_to_text(mv.get("outlook_summary"))
-    pos = html_to_text(mv.get("company_positioning"))
-    dist = mv.get("outlook_distribution") or []
-    dist2 = mv.get("ratings_distribution") or []
-
-    if not bnd_l or not bnd_r:
-        title_bot = title_bottom(s, default=Inches(1.1), text=title_text)
-        fallback_top = max(title_bot + Inches(0.2), Inches(1.3))
-        fallback_h = Inches(5.5)
-        if not bnd_l:
-            bnd_l = (Inches(0.6), fallback_top, Inches(6.0), fallback_h)
-        if not bnd_r:
-            bnd_r = (Inches(6.9), fallback_top, Inches(6.0), fallback_h)
-
-    if summary or pos:
-        l, t, w, h = bnd_l
-        tb = s.shapes.add_textbox(l, t, w, h)
-        tf = tb.text_frame
-        tf.word_wrap = True
-        tf.auto_size = MSO_AUTO_SIZE.NONE
-        tf.text = ""
-        lines = [x for x in [summary, pos] if x]
-        for i, line in enumerate(lines):
-            p = tf.paragraphs[0] if i == 0 else tf.add_paragraph()
-            p.line_spacing = 1.2
-            p.space_after = Pt(6)
-            r = p.add_run()
-            r.text = line
-            r.font.name = "Arial"
-            r.font.size = Pt(11)
-            r.font.color.rgb = TEXT_DARK
-
-    pie_bounds = bar_bounds = None
-    if dist and dist2:
-        l, t, w, h = bnd_r
-        half = int(h / 2)
-        pie_bounds = (l, t, w, half)
-        bar_bounds = (l, t + half, w, h - half)
-    elif dist:
-        pie_bounds = bnd_r
-    elif dist2:
-        bar_bounds = bnd_r
-
-    if pie_bounds and dist:
-        l, t, w, h = pie_bounds
-        labels, values, colors = [], [], []
-        for d in dist:
-            cat = str(d.get("category", ""))
-            pct = d.get("percentage", d.get("count", 0))
-            count = d.get("count", pct or 0)
-            try:
-                if count is None or float(count) <= 0:
-                    continue
-            except (TypeError, ValueError):
-                continue
-            labels.append(cat)
-            values.append(count)
-            colors.append(OUTLOOK_PALETTE.get(cat.lower(), BRIGHT_BLUE))
-        if values:
-            add_pie_chart(s, l, t, w, h,
-                          "Moody's Outlooks Distribution",
-                          labels, values, colors=colors, use_legend=True)
-
-    if bar_bounds and dist2:
-        l, t, w, h = bar_bounds
-        add_bar_chart(s, l, t, w, h,
-                      "Moody's Ratings Distribution",
-                      [str(d.get("rating", "")) for d in dist2],
-                      [d.get("count", 0) for d in dist2],
-                      color=BRIGHT_BLUE)
-    return s
-
-
-def build_macro_outlook(prs, data):
-    s = prs.slides.add_slide(prs.slide_layouts[L_ONE_COL])
-    set_text(s, 0, "Global Macro Outlook")
-    set_text(s, 18, "MOODY'S REAL GDP GROWTH")
-    macro = data.get("sections", {}).get("macro_outlook", {}) or {}
-    set_text(s, 29, "")
-    set_text(s, 19, "Source: Moodys.com")
-    # Replace the content placeholder with bullets above a GDP table.
-    bnd = ph_bounds(s, 22)
-    del_ph(s, 22)
-    if not bnd:
-        return s
-    l, t, w, h = bnd
-    new_t = Inches(1.65)
-    h = h + (t - new_t)
-    t = new_t
-    bullet_h = Inches(1.2)
-    add_textbox_bullets(s, l, t, w, bullet_h,
-                        macro.get("gdp_commentary", []), size=10)
-    tbl = macro.get("gdp_table", {}) or {}
-    years = tbl.get("year_columns", []) or []
-    rows = tbl.get("rows", []) or []
-    if years and rows:
-        headers = ["Country"] + [f"{y} GDP Growth" for y in years]
-        body = [[r.get("country", "")] + list(r.get("values", []))
-                for r in rows]
-        add_table(s, l, t + bullet_h, w, h - bullet_h, headers, body,
-                  col_widths=[2.0] + [1.2] * len(years),
-                  header_size=10, body_size=9)
-    return s
-
-
-def _fit_bullet_font(width_emu, height_emu, items, size, bullet="•"):
-    """Pick a font size (<= ``size``, >= 8) that makes the bullet list fit
-    vertically within ``height_emu``. Uses a conservative chars-per-line
-    estimate for Arial and iterates 1pt at a time."""
-    # Coerce bare string so we don't iterate characters
-    if isinstance(items, str):
-        items = [items] if items.strip() else []
-    try:
-        width_pt = (float(width_emu) / Inches(1)) * 72.0
-        avail_pt = (float(height_emu) / Inches(1)) * 72.0
-    except Exception:
-        return size
-    if width_pt <= 0 or avail_pt <= 0 or not items:
-        return size
-    pad_pt = 14.4
-    for sz in range(int(size), 7, -1):
-        avg_char_pt = sz * 0.52
-        effective_pt = max(1.0, width_pt - pad_pt)
-        cpl = max(10, int(effective_pt / avg_char_pt))
-        total_lines = 0
-        for item in items:
-            text = f"{bullet} {item}"
-            total_lines += max(1, -(-len(text) // cpl))
-        line_spacing = 1.2 if sz <= 9 else 1.25
-        space_after = 2 if sz <= 9 else 4
-        used_pt = total_lines * sz * line_spacing + len(items) * space_after
-        if used_pt <= avail_pt * 0.95:
-            return sz
-    return 8
-
-
-def add_textbox_bullets(slide, left, top, width, height, items, *,
-                        size=11, bullet="•"):
-    tb = slide.shapes.add_textbox(left, top, width, height)
-    tf = tb.text_frame
-    tf.word_wrap = True
-    tf.text = ""
-    # Guard: a bare string must be wrapped in a list; iterating a string
-    # directly produces one character per bullet (the "• C \n • o \n …" bug).
-    if isinstance(items, str):
-        items = [items] if items.strip() else []
-    items = [str(x) for x in (items or []) if x is not None and str(x).strip()]
-    if not items:
-        return tb
-    fit_size = _fit_bullet_font(width, height, items, size, bullet=bullet)
-    line_spacing = 1.2 if fit_size <= 9 else 1.25
-    space_after_pt = 2 if fit_size <= 9 else 4
-    for i, item in enumerate(items):
-        p = tf.paragraphs[0] if i == 0 else tf.add_paragraph()
-        p.alignment = PP_ALIGN.LEFT
-        p.line_spacing = line_spacing
-        p.space_after = Pt(space_after_pt)
-        r0 = p.add_run()
-        r0.text = f"{bullet} "
-        r0.font.name = "Arial"
-        r0.font.size = Pt(fit_size)
-        r0.font.color.rgb = TEXT_DARK
-        r1 = p.add_run()
-        r1.text = item
-        r1.font.name = "Arial"
-        r1.font.size = Pt(fit_size)
-        r1.font.color.rgb = TEXT_DARK
-    return tb
-
-
-def build_rating_actions(prs, data):
-    s = prs.slides.add_slide(prs.slide_layouts[L_ONE_COL])
-    set_text(s, 0, "Moody's Rating Actions – YTD")
-    set_text(s, 18, "")
-    set_text(s, 29, "")
-    set_text(s, 19, "Source: Moodys.com")
-    actions = data.get("sections", {}).get("rating_actions_ytd", []) or []
-    items = []
-    for a in actions[:10]:
-        date = a.get("date", "")
-        company = a.get("company", "")
-        summary = a.get("summary", "")
-        prefix = f"{date} – {company}" if company else date
-        items.append({"bold": prefix, "text": summary})
-    set_bold_prefix_bullets(s, 22, items, size=11)
-    return s
-
-
-def build_financial_analysis(prs, data):
-    s = prs.slides.add_slide(prs.slide_layouts[L_TWO_COL])
-    company = data.get("target_company", "")
-    set_text(s, 0, f"{company} financial analysis")
-    set_text(s, 21, "")
-    set_text(s, 29, "")
-    set_text(s, 30, "")
-    set_text(s, 19, "Source: Moodys.com, Company filings")
-    fa = data.get("sections", {}).get("financial_analysis", {}) or {}
-    # Left column (27): bullets.
-    bnd_l = ph_bounds(s, 27)
-    bnd_r = ph_bounds(s, 28)
-    del_ph(s, 27)
-    del_ph(s, 28)
-    new_t = Inches(1.15)
-    if bnd_l:
-        l, t, w, h = bnd_l
-        h = h + (t - new_t)
-        t = new_t
-        _fa_commentary = fa.get("commentary", [])
-        if isinstance(_fa_commentary, str):
-            _fa_commentary = [_fa_commentary] if _fa_commentary.strip() else []
-        add_textbox_bullets(s, l, t, w, h, _fa_commentary, size=10)
-    if bnd_r:
-        l, t, w, h = bnd_r
-        h = h + (t - new_t)
-        t = new_t
-        history = fa.get("rating_history", []) or []
-        # ── Sort oldest → newest so chart x-axis reads left-to-right chronologically ──
-        _MONTH_MAP = {
-            "jan":1,"feb":2,"mar":3,"apr":4,"may":5,"jun":6,
-            "jul":7,"aug":8,"sep":9,"oct":10,"nov":11,"dec":12
-        }
-        def _date_key(entry: dict) -> tuple:
-            """Return (year, month) tuple for sorting; fall back to (0,0) if unparseable."""
-            raw = entry.get("date", "")
-            m = __import__("re").search(r"(\d{4})", raw)
-            year = int(m.group(1)) if m else 0
-            mn = __import__("re").search(r"(jan|feb|mar|apr|may|jun|jul|aug|sep|oct|nov|dec)", raw.lower())
-            month = _MONTH_MAP.get(mn.group(1), 0) if mn else 0
-            return (year, month)
-        history = sorted(history, key=_date_key)
-        # Rebuild chart_data in the same sorted order (parallel arrays must stay aligned)
-        _notch = {"Aaa":21,"Aa1":20,"Aa2":19,"Aa3":18,"A1":17,"A2":16,"A3":15,
-                  "Baa1":14,"Baa2":13,"Baa3":12,"Ba1":11,"Ba2":10,"Ba3":9,
-                  "B1":8,"B2":7,"B3":6,"Caa1":5,"Caa2":4,"Caa3":3,"Ca":2,"C":1}
-        chart_data = [_notch.get(entry.get("rating",""), 0) for entry in history] if history else (fa.get("rating_chart_data", []) or [])
-        chart_h = int(h * 0.5)
-        if chart_data:
-            cats = [entry.get("date", f"P{i+1}") for i, entry in enumerate(history)]
-            if len(cats) < len(chart_data):
-                cats += [f"P{i+1}" for i in range(len(cats), len(chart_data))]
-            elif len(cats) > len(chart_data):
-                cats = cats[:len(chart_data)]
-            # Single data point: pad with today so the line has direction
-            # and reflects that the rating is still current as of report date
-            if len(chart_data) == 1:
-                import datetime
-                today_label = datetime.date.today().strftime("%b %Y")
-                cats        = cats + [today_label]
-                chart_data  = chart_data + [chart_data[0]]
-            add_line_chart(s, l, t, w, chart_h,
-                           f"Moody's Rating History for {company}",
-                           cats, chart_data, color=BRIGHT_BLUE,
-                           rating_scale=True)
-        rationale_top = t + chart_h + Inches(0.15)
-        add_textbox(s, l, rationale_top, w, Inches(0.3),
-                    "Moody's rating rationale", size=11, bold=True,
-                    color=TEXT_DARK)
-        items = []
-        for h2 in history:
-            items.append({
-                "bold": h2.get("date", ""),
-                "text": h2.get("reason", h2.get("rating", "")),
-            })
-        _add_bold_prefix_textbox(s, l, rationale_top + Inches(0.35), w,
-                         Inches(2.4), items, size=10)
-    return s
-
-
-def _add_bold_prefix_textbox(slide, left, top, width, height, items, *,
-                             size=10, bullet="•"):
-    tb = slide.shapes.add_textbox(left, top, width, height)
-    tf = tb.text_frame
-    tf.word_wrap = True
-    tf.text = ""
-    for i, item in enumerate(items or []):
-        if isinstance(item, dict):
-            prefix = item.get("bold", "")
-            rest = item.get("text", "")
-        else:
-            prefix, rest = "", str(item)
-        p = tf.paragraphs[0] if i == 0 else tf.add_paragraph()
-        p.alignment = PP_ALIGN.LEFT
-        p.line_spacing = 1.2
-        p.space_after = Pt(3)
-        r0 = p.add_run()
-        r0.text = f"{bullet} "
-        r0.font.name = "Arial"
-        r0.font.size = Pt(size)
-        r0.font.color.rgb = TEXT_DARK
-        if prefix:
-            r1 = p.add_run()
-            r1.text = f"{prefix}: "
-            r1.font.name = "Arial"
-            r1.font.size = Pt(size)
-            r1.font.bold = True
-            r1.font.color.rgb = TEXT_DARK
-        if rest:
-            r2 = p.add_run()
-            r2.text = rest
-            r2.font.name = "Arial"
-            r2.font.size = Pt(size)
-            r2.font.color.rgb = TEXT_DARK
-    return tb
-
-
-def build_revenue_distribution(prs, data):
-    s = prs.slides.add_slide(prs.slide_layouts[L_ONE_COL])
-    company = data.get("target_company", "")
-    set_text(s, 0, "Revenue distribution")
-    set_text(s, 18, "")
-    set_text(s, 29, "")
-    set_text(s, 19, "Source: Moodys.com, Company filings")
-    rd = data.get("sections", {}).get("revenue_distribution", {}) or {}
-    bnd = ph_bounds(s, 22)
-    del_ph(s, 22)
-    if not bnd:
-        return s
-    l, t, w, h = bnd
-    new_t = Inches(1.15)      # move up from 2.653"
-    h = h + (t - new_t)      # recover ~1.5" into h
-    t = new_t
-
-    bullet_h = Inches(0.6)    # was 1.2 — only 2 short bullets, 0.6" is enough
-    add_textbox_bullets(s, l, t, w, bullet_h,
-                        rd.get("commentary", []), size=10)
-    geo = rd.get("by_geography", []) or []
-    seg = rd.get("by_segment", []) or []
-    charts_top = t + bullet_h   # shifts up automatically
-    charts_h = h - bullet_h     # charts get all the recovered space
-    col_w = int((w - Inches(0.4)) / 2)
-    def _clean_slices(items):
-        out = []
-        for it in items or []:
-            name = str(it.get("name", "")).strip()
-            try:
-                pct = float(it.get("percentage", 0) or 0)
-            except (TypeError, ValueError):
-                pct = 0.0
-            if not name or pct <= 0:
-                continue
-            out.append((name, pct))
-        return out
-
-    geo_clean = _clean_slices(geo)
-    seg_clean = _clean_slices(seg)
-
-    if geo_clean:
-        add_pie_chart(s, l, charts_top, col_w, charts_h,
-                      f"{company} Revenue by region",
-                      [n for n, _ in geo_clean],
-                      [p for _, p in geo_clean],
-                      use_legend=True)
-    if seg_clean:
-        add_pie_chart(s, l + col_w + Inches(0.4), charts_top, col_w,
-                      charts_h, f"{company} Revenue by segment",
-                      [n for n, _ in seg_clean],
-                      [p for _, p in seg_clean],
-                      use_legend=True)
-    return s
-
-
-def build_swot(prs, data):
-    s = prs.slides.add_slide(prs.slide_layouts[L_ONE_COL])
-    set_text(s, 0, "SWOT Analysis")
-    set_text(s, 18, "")
-    set_text(s, 29, "")
-    set_text(s, 19, "Source: Moodys.com, Company filings")
-    swot = data.get("sections", {}).get("swot", {}) or {}
-    bnd = ph_bounds(s, 22)
-    del_ph(s, 22)
-    if not bnd:
-        return s
-    l, t, w, h = bnd
-    new_t = Inches(1.15)       
-    h = h + (t - new_t)    
-    t = new_t
-    col_w = int((w - Inches(0.4)) / 2)
-    row_h = int((h - Inches(0.2)) / 2)
-    quads = [
-        ("Strengths", swot.get("strengths", [])),
-        ("Weaknesses", swot.get("weaknesses", [])),
-        ("Opportunities", swot.get("opportunities", [])),
-        ("Threats", swot.get("threats", [])),
-    ]
-    for i, (label, items) in enumerate(quads):
-        col = i % 2
-        row = i // 2
-        x = l + col * (col_w + Inches(0.4))
-        y = t + row * (row_h + Inches(0.2))
-        add_textbox(s, x, y, col_w, Inches(0.3), label,
-                    size=11, bold=True, color=TEXT_DARK)
-        add_textbox_bullets(s, x, y + Inches(0.3), col_w,
-                            row_h - Inches(0.3), items, size=10)
-    return s
-
-
-def build_key_metrics(prs, data):
-    s = prs.slides.add_slide(prs.slide_layouts[L_ONE_COL])
-    company = data.get("target_company", "")
-    set_text(s, 0, f"Key financial metrics - {company}")
-    set_text(s, 18, "")
-    set_text(s, 29, "")
-    set_text(s, 19, "Source: Moodys.com, Company filings")
-    km = data.get("sections", {}).get("key_metrics", {}) or {}
-    periods = km.get("periods", []) or []
-    charts = [
-        ("Revenue", km.get("revenue", []), "#,##0"),
-        ("EBIT Margin", km.get("ebit_margin", []), "0.0\"%\""),
-        ("Debt / EBITDA", km.get("debt_ebitda", []), "0.0\"x\""),
-        ("RCF / Net Debt", km.get("rcf_net_debt", []), "0.0\"%\""),
-    ]
-    bnd = ph_bounds(s, 22)
-    del_ph(s, 22)
-    if not bnd:
-        return s
-    l, t, w, h = bnd
-    new_t = Inches(1.15)
-    h = h + (t - new_t)
-    t = new_t
-
-    gap = Inches(0.3)
-    col_w = int((w - gap) / 2)
-    row_h = int((h - gap) / 2)    # ~1.5" taller rows → charts fill the slide properly
-    for i, (title, vals, fmt) in enumerate(charts):
-        col = i % 2
-        row = i // 2
-        x = l + col * (col_w + gap)
-        y = t + row * (row_h + gap)
-        cats = periods[:len(vals)] if periods else [f"P{k+1}" for k in range(len(vals))]
-        if vals:
-            add_bar_chart(s, x, y, col_w, row_h, title, cats, vals,
-                          color=BRIGHT_BLUE, y_format=fmt)
-        else:
-            add_textbox(s, x, y, col_w, row_h, f"{title}\n(no data)",
-                        size=11, color=TEXT_MUTED, align=PP_ALIGN.CENTER,
-                        anchor=MSO_ANCHOR.MIDDLE)
-    return s
-
-
-def build_strategic_updates(prs, data):
-    s = prs.slides.add_slide(prs.slide_layouts[L_TWO_COL])
-    company = data.get("target_company", "")
-    set_text(s, 0, f"{company} strategic updates")
-    set_text(s, 21, "")
-    set_text(s, 29, "RECENT DEVELOPMENTS")
-    set_text(s, 30, "FORWARD LOOKING")
-    set_text(s, 19, "Source: Moodys.com, Company filings")
-    sup = data.get("sections", {}).get("strategic_updates", {}) or {}
-    set_bullets(s, 27, sup.get("recent", []), size=11)
-    set_bullets(s, 28, sup.get("forward", []), size=11)
-    return s
-
-
-def build_categorized(prs, data, section_key, title, source):
-    s = prs.slides.add_slide(prs.slide_layouts[L_ONE_COL])
-    set_text(s, 0, title)
-    set_text(s, 18, "")
-    set_text(s, 29, "")
-    set_text(s, 19, f"Source: {source}")
-    groups = _normalize_categorized(data.get("sections", {}).get(section_key))
-    bnd = ph_bounds(s, 22)
-    del_ph(s, 22)
-    if not bnd or not groups:
-        return s
-    l, t, w, h = bnd
-    new_t = Inches(1.15)
-    h = h + (t - new_t)
-    t = new_t
-
-    per_h = int(h / max(len(groups), 1))   # each group gets a proportionally larger slice
-    for i, g in enumerate(groups):
-        y = t + i * per_h
-        add_textbox(s, l, y, w, Inches(0.3), g.get("name", ""),
-                    size=11, bold=True, color=TEXT_DARK)
-        items = g.get("items") or []
-        bullets = []
-        for it in items:
-            if isinstance(it, dict):
-                prefix = it.get("date") or it.get("when") or ""
-                body = it.get("text") or it.get("summary") or ""
-                bullets.append({"bold": f"{prefix} →" if prefix else "",
-                                "text": body})
-            else:
-                bullets.append(str(it))
-        if any(isinstance(b, dict) for b in bullets):
-            _add_bold_prefix_textbox(s, l, y + Inches(0.3), w,
-                                     per_h - Inches(0.3),
-                                     bullets, size=10)
-        else:
-            add_textbox_bullets(s, l, y + Inches(0.3), w,
-                                per_h - Inches(0.3), bullets, size=10)
-    return s
-
-
-def build_peer_summary(prs, data):
-    s = prs.slides.add_slide(prs.slide_layouts[L_ONE_COL])
-    set_text(s, 0, "Peer comparison summary")
-    set_text(s, 18, "")
-    set_text(s, 29, "")
-    set_text(s, 19, "Source: Moodys.com, Company filings")
-    ps = data.get("sections", {}).get("peer_summary", {}) or {}
-    table = ps.get("table", []) or []
-    bnd = ph_bounds(s, 22)
-    del_ph(s, 22)
-    if not bnd:
-        return s
-    l, t, w, h = bnd
-    new_t = Inches(1.15)
-    h = h + (t - new_t)
-    t = new_t
-
-    bul_h = Inches(0.9)
-    _ps_commentary = ps.get("commentary", [])
-    if isinstance(_ps_commentary, str):
-        _ps_commentary = [_ps_commentary] if _ps_commentary.strip() else []
-    add_textbox_bullets(s, l, t, w, bul_h, _ps_commentary, size=10)
-    if not table:
-        return s
-    fields = [
-        ("Country", "country"),
-        ("Market Capitalization", "market_cap"),
-        ("Moody's Ratings", "rating"),
-        ("Outlook", "outlook"),
-        ("Business Mix", "business_mix"),
-    ]
-    headers = [""] + [c.get("company", "") for c in table]
-    body = []
-    for label, key in fields:
-        body.append([label] + [str(c.get(key, "") or "") for c in table])
-    tbl_top = t + bul_h
-    tbl_h = h - bul_h
-    add_table(s, l, tbl_top, w, tbl_h, headers, body,
-              first_col_bold=True,
-              col_widths=[1.6] + [1.0] * len(table),
-              header_size=10, body_size=9)
-    return s
-
-
-def build_peer_financials(prs, data):
-    s = prs.slides.add_slide(prs.slide_layouts[L_ONE_COL])
-    set_text(s, 0, "Detailed peer comparison")
-    set_text(s, 18, "")
-    set_text(s, 29, "")
-    set_text(s, 19, "Source: Moodys.com, Company filings")
-    pf = data.get("sections", {}).get("peer_financials", {}) or {}
-    columns = pf.get("columns", []) or []
-    rows = pf.get("rows", []) or []
-    bnd = ph_bounds(s, 22)
-    del_ph(s, 22)
-    if not bnd or not rows:
-        return s
-    l, t, w, h = bnd
-
-    # Center the table vertically between title (ends ~1.08") and footer (6.583")
-    tbl_h = Inches(3.8)
-    avail_top = Inches(1.15)
-    avail_bot = Inches(6.50)
-    t = avail_top + int((avail_bot - avail_top - tbl_h) / 2)  # center it
-
-    headers = ["Financial Items (in millions)"] + [r.get("company", "") for r in rows]
-    body = [["Period"] + [str(r.get("period", "") or "") for r in rows],
-            ["Currency"] + [str(r.get("currency", "") or "") for r in rows]]
-    for i, label in enumerate(columns):
-        body.append(
-            [label] + [str((r.get("values") or [""] * len(columns))[i]
-                       if i < len(r.get("values", [])) else "")
-                       for r in rows]
-        )
-    add_table(s, l, t, w, tbl_h, headers, body,
-              first_col_bold=True,
-              col_widths=[2.0] + [1.0] * len(rows),
-              header_size=10, body_size=9)
-    return s
-
-
-def _build_peer_bars(prs, data, section_key, title, chart_defs):
-    s = prs.slides.add_slide(prs.slide_layouts[L_ONE_COL])
-    set_text(s, 0, title)
-    set_text(s, 18, "")
-    set_text(s, 29, "")
-    set_text(s, 19, "Source: Moodys.com, Company filings")
-    section = data.get("sections", {}).get(section_key, {}) or {}
-    bnd = ph_bounds(s, 22)
-    del_ph(s, 22)
-    if not bnd:
-        return s
-    l, t, w, h = bnd
-
-    new_t = Inches(1.15)
-    h = h + (t - new_t)
-    t = new_t
-
-    commentary = section.get("commentary")
-    if isinstance(commentary, str):
-        commentary = [commentary]
-    bul_h = Inches(0.6)          # was 1.0 — commentary is 1-2 lines, 0.6" is enough
-    add_textbox_bullets(s, l, t, w, bul_h, commentary or [], size=10)
-    charts_top = t + bul_h       # shifts up automatically
-    charts_h = h - bul_h        # charts get ~1.9" more height total
-    col_w = int((w - Inches(0.4)) / 2)
-    for i, (chart_title, key, fmt) in enumerate(chart_defs):
-        x = l + i * (col_w + Inches(0.4))
-        entries = section.get(key, []) or []
-        clean = [(e.get("company", ""), e.get("value"), e.get("period", ""))
-                 for e in entries if e.get("value") is not None]
-        if not clean:
-            continue
-        clean.sort(key=lambda p: (p[1] if p[1] is not None else -1e9))
-        periods_present = [c[2] for c in clean if c[2]]
-        all_same_period = len(set(periods_present)) <= 1
-        if all_same_period:
-            cat_labels = [c[0] for c in clean]
-        else:
-            cat_labels = [f"{c[0]} ({c[2]})" if c[2] else c[0] for c in clean]
-        add_bar_chart(s, x, charts_top, col_w, charts_h, chart_title,
-                      cat_labels, [c[1] for c in clean],
-                      color=BRIGHT_BLUE, y_format=fmt)
-    return s
-
-
-def build_peer_debt(prs, data):
-    return _build_peer_bars(prs, data, "peer_debt_charts",
-                            "Peer comparison – Debt",
-                            [("Debt / EBITDA", "debt_ebitda", "0.0\"x\""),
-                             ("RCF / Net Debt", "rcf_net_debt", "0.0\"%\"")])
-
-
-def build_peer_profitability(prs, data):
-    return _build_peer_bars(prs, data, "peer_profitability_charts",
-                            "Peer comparison – Profitability",
-                            [("Revenue", "revenue", "#,##0"),
-                             ("EBIT Margin", "ebit_margin", "0.0\"%\"")])
-
-
-def build_peer_scatter(prs, data):
-    s = prs.slides.add_slide(prs.slide_layouts[L_ONE_COL])
-    set_text(s, 0, "Peer comparison")
-    set_text(s, 18, "")
-    set_text(s, 29, "")
-    set_text(s, 19, "Source: Moodys.com, Company filings")
-    ps = data.get("sections", {}).get("peer_scatter", {}) or {}
-    bnd = ph_bounds(s, 22)
-    del_ph(s, 22)
-    if not bnd:
-        return s
-    l, t, w, h = bnd
-
-    new_t = Inches(1.15)
-    h = h + (t - new_t)
-    t = new_t
-
-    commentary = ps.get("commentary")
-    if isinstance(commentary, str):
-        commentary = [commentary]
-    bul_h = Inches(0.6)          # was 1.0
-    add_textbox_bullets(s, l, t, w, bul_h, commentary or [], size=10)
-    charts_top = t + bul_h
-    charts_h = h - bul_h
-    col_w = int((w - Inches(0.4)) / 2)
-    sc1 = ps.get("margin_vs_leverage", []) or []
-    sc2 = ps.get("fcf_vs_rcf", []) or []
-    _, c1 = add_scatter_chart(s, l, charts_top, col_w, charts_h,
-                              "EBIT Margin vs Debt/EBITDA", sc1,
-                              x_label="Debt/EBITDA",
-                              y_label="EBIT Margin (%)")
-    if c1:
-        add_scatter_labels(s, l, charts_top, col_w, charts_h, c1)
-    x2 = l + col_w + Inches(0.4)
-    _, c2 = add_scatter_chart(s, x2, charts_top, col_w, charts_h,
-                              "FCF/Debt vs RCF / Debt", sc2,
-                              x_label="RCF / Debt", y_label="FCF / Debt")
-    if c2:
-        add_scatter_labels(s, x2, charts_top, col_w, charts_h, c2)
-    return s
-
-
-def build_scorecard(prs, data):
-    s = prs.slides.add_slide(prs.slide_layouts[L_ONE_COL])
-    set_text(s, 0, "Peer Scorecard Outcome Comparison")
-    set_text(s, 18, "")
-    set_text(s, 29, "")
-    set_text(s, 19, "Source: Moodys.com")
-    sc = data.get("sections", {}).get("scorecard", {}) or {}
-    companies = sc.get("companies", []) or []
-    factors = sc.get("factors", []) or []
-    is_header = sc.get("is_header", []) or []
-    values = sc.get("values", []) or []
-    bnd = ph_bounds(s, 22)
-    del_ph(s, 22)
-    if not bnd or not companies or not factors:
-        return s
-    l, _t, w, _h = bnd
-    t = Inches(1.2)
-    h = Inches(5.85)
-    headers = ["Factors"]
-    for name in companies:
-        headers.append(name)
-        headers.append("")
-    sub_header = [""] + ["Measure", "Score"] * len(companies)
-    body = [sub_header]
-    header_rows = []
-    for idx, label in enumerate(factors):
-        hdr = is_header[idx] if idx < len(is_header) else False
-        row = [label]
-        rv = values[idx] if idx < len(values) else []
-        for i in range(len(companies)):
-            pair = rv[i + 1] if i + 1 < len(rv) else []
-            measure = pair[0] if len(pair) > 0 else ""
-            score = pair[1] if len(pair) > 1 else ""
-            row.append(str(measure))
-            row.append(str(score))
-        body.append(row)
-        if hdr:
-            header_rows.append(len(body) - 1)
-    nrows_total = len(body) + 1
-    if nrows_total <= 16:
-        body_size, header_size = 9, 10
-    elif nrows_total <= 19:
-        body_size, header_size = 8, 9
-    else:
-        body_size, header_size = 7, 8
-    table = add_table(s, l, t, w, h, headers, body,
-                      first_col_bold=True,
-                      col_widths=[2.4] + [0.65] * (len(headers) - 1),
-                      header_size=header_size, body_size=body_size)
-    tight_top = Emu(18000)
-    tight_bot = Emu(18000)
-    tight_left = Emu(36000)
-    tight_right = Emu(36000)
-    for r in table.rows:
-        for c in r.cells:
-            try:
-                c.margin_top = tight_top
-                c.margin_bottom = tight_bot
-                c.margin_left = tight_left
-                c.margin_right = tight_right
-            except Exception:
-                pass
-    try:
-        per_row = int(h / nrows_total)
-        for r in table.rows:
-            r.height = per_row
-    except Exception:
-        pass
-    for row_idx in header_rows:
-        for j in range(len(headers)):
-            c = table.cell(row_idx, j)
-            c.fill.solid()
-            c.fill.fore_color.rgb = HEADER_FILL
-            for p in c.text_frame.paragraphs:
-                for r in p.runs:
-                    r.font.bold = True
-    return s
-
-
-def build_esg(prs, data):
-    s = prs.slides.add_slide(prs.slide_layouts[L_TWO_COL])
-    company = data.get("target_company", "")
-    set_text(s, 0, "ESG Analysis")
-    set_text(s, 21, "")
-    set_text(s, 29, "ESG TABLE")
-    set_text(s, 30, f"{company.upper()} ESG")
-    set_text(s, 19, "Source: Moodys.com, Company filings")
-    esg = data.get("sections", {}).get("esg_analysis", {}) or {}
-    rows = esg.get("table", []) or []
-    bnd_l = ph_bounds(s, 27)
-    bnd_r = ph_bounds(s, 28)
-    del_ph(s, 27)
-    del_ph(s, 28)
-    if bnd_l and rows:
-        l, t, w, h = bnd_l
-        headers = ["Company", "CIS", "Env.", "Social", "Gov."]
-        body = [[r.get("company", ""),
-                 r.get("cis", ""),
-                 r.get("environmental", ""),
-                 r.get("social", ""),
-                 r.get("governance", "")] for r in rows]
-        add_table(s, l, t, w, min(h, Inches(0.5) + Inches(0.45) * len(body)),
-                  headers, body, first_col_bold=True,
-                  col_widths=[1.5, 1.0, 0.9, 0.9, 0.9],
-                  header_size=10, body_size=9)
-    if bnd_r:
-        l, t, w, h = bnd_r
-        _esg_commentary = esg.get("commentary", [])
-        if isinstance(_esg_commentary, str):
-            _esg_commentary = [_esg_commentary] if _esg_commentary.strip() else []
-        add_textbox_bullets(s, l, t, w, h, _esg_commentary, size=11)
-    return s
-
-
-def build_thank_you(prs, data):
-    s = prs.slides.add_slide(prs.slide_layouts[L_BACK_COVER])
-    set_text(s, 0, "Thank you")
-    set_text(s, 94, "MOODY'S ANALYTICS")
-    set_text(s, 95, data.get("sector", ""))
-    set_text(s, 96, data.get("report_date", ""))
-    return s
-
-
-def build_disclaimer(prs):
-    s = prs.slides.add_slide(prs.slide_layouts[L_DISCLAIMER])
-    return s
-
-
-# ---- Orchestration ---------------------------------------------------------
-
-def build(payload: dict, out_path: Path) -> None:
-    prs = Presentation(str(TEMPLATE_PATH))
-    remove_all_slides(prs)
-
-    build_cover(prs, payload)                                           # 1
-    build_agenda(prs, payload)                                          # 2
-
-    build_divider(prs, "Sector Analysis")                               # 3
-    build_sector_overview(prs, payload)                                 # 4
-    build_moodys_view(prs, payload)                                     # 5
-    build_macro_outlook(prs, payload)                                   # 6
-    build_rating_actions(prs, payload)                                  # 7
-
-    build_divider(prs, "Company Credit Overview")                       # 8
-    build_financial_analysis(prs, payload)                              # 9
-    build_revenue_distribution(prs, payload)                            # 10
-    build_swot(prs, payload)                                            # 11
-    build_key_metrics(prs, payload)                                     # 12
-    build_strategic_updates(prs, payload)                               # 13
-    build_categorized(prs, payload, "news_mna",
-                      "News/M&A/Leadership", "Web search")              # 14
-    build_categorized(prs, payload, "external_trends",
-                      "External trends / pressures / reputational risks",
-                      "Web search")                                     # 15
-
-    build_divider(prs, "Company Positioning vs. Peers")                 # 16
-    build_peer_summary(prs, payload)                                    # 17
-    build_peer_financials(prs, payload)                                 # 18
-    build_peer_debt(prs, payload)                                       # 19
-    build_peer_profitability(prs, payload)                              # 20
-    build_peer_scatter(prs, payload)                                    # 21
-    build_scorecard(prs, payload)                                       # 22
-    build_esg(prs, payload)                                             # 23
-
-    build_thank_you(prs, payload)                                       # 24
-    build_disclaimer(prs)                                               # 25
-
-    for slide in prs.slides:
-        clear_empty_placeholders(slide)
-
-    out_path.parent.mkdir(parents=True, exist_ok=True)
-    prs.save(str(out_path))
-
-
-def main() -> int:
-    if len(sys.argv) != 3:
-        print("Usage: build_pptx.py <payload.json> <output.pptx>",
-              file=sys.stderr)
-        return 2
-    with Path(sys.argv[1]).open("r", encoding="utf-8") as f:
-        payload = json.load(f)
-    build(payload, Path(sys.argv[2]))
-    print(f"Wrote {sys.argv[2]}")
-    return 0
-
-
-if __name__ == "__main__":
-    sys.exit(main())
diff --git a/plugins/moody-s/skills/moody-s-rating-analysis/scripts/requirements.txt b/plugins/moody-s/skills/moody-s-rating-analysis/scripts/requirements.txt
index 529f7b45..f35e5c50 100644
--- a/plugins/moody-s/skills/moody-s-rating-analysis/scripts/requirements.txt
+++ b/plugins/moody-s/skills/moody-s-rating-analysis/scripts/requirements.txt
@@ -1 +1,3 @@
-python-pptx>=0.6.21
+# No additional Python dependencies required.
+# build_html.py uses only the Python standard library.
+# Charts are rendered in-browser via Chart.js (loaded from CDN).
diff --git a/plugins/moody-s/skills/moody-s-sector-brief/SKILL.md b/plugins/moody-s/skills/moody-s-sector-brief/SKILL.md
index c8482026..5f5730ee 100644
--- a/plugins/moody-s/skills/moody-s-sector-brief/SKILL.md
+++ b/plugins/moody-s/skills/moody-s-sector-brief/SKILL.md
@@ -23,21 +23,11 @@ client-side render logic. The fenced code block is the deliverable.
 
 > ## ⚠️ CRITICAL — NON-NEGOTIABLE OUTPUT CONTRACT
 >
-> **The LLM MUST stream the final report back as a single HTML artifact inside the assistant
-> response.** This is the only acceptable form of delivery for this skill. Specifically:
+> **The LLM MUST deliver the report as a single HTML artifact in the assistant response.** Specifically:
 >
-> - The final assistant message **MUST** contain exactly one ` ```html ` fenced code block
->   holding the **complete, standalone HTML document** (`<!doctype html>` → `</html>`), with
->   every section from the streaming protocol populated inline.
-> - The LLM **MUST NOT** write the report to a file on disk (no `Write`, no `cp` of the
->   template, no `StrReplace` into a working artifact, no `open` command).
-> - The LLM **MUST NOT** split the report across multiple code blocks, multiple messages,
->   partial snippets, or summaries.
-> - The LLM **MUST NOT** substitute prose, Markdown, JSON, attachments, or links for the
->   fenced HTML artifact. The artifact itself is the answer.
-> - If data gathering fails partially, still emit the single ` ```html ` artifact with
->   the best-available content and brief `"Data unavailable"` placeholders for missing
->   sections — never skip the artifact.
+> - The final message **MUST** contain exactly one ` ```html ` fenced code block holding the **complete, standalone HTML document** (`<!doctype html>` → `</html>`), every section populated inline.
+> - The LLM **MUST NOT** write to disk, split across messages/blocks, or substitute prose, Markdown, JSON, or links for the artifact.
+> - If data partially fails, emit the artifact with `"Data unavailable"` placeholders — never skip it.
 >
 > Treat any other output shape as a hard failure of the skill.
 
@@ -68,27 +58,9 @@ Before emitting the HTML report, **read both**:
    (inside `<template>` tags) for the document head, cover, TOC, section block, sources-section
    wrapper, footer, and outlook-badge.
 
-**Lookup order — always check the shared template before inventing.** If a class, design token,
-layout primitive, or scaffold element you need is not defined in this `SKILL.md` or already
-present in this skill's `assets/template.html`, the shared template skill is authoritative. Do
-not invent CSS, HTML scaffolds, or design tokens that the shared skill already provides; do not
-silently restyle anything the shared skill owns (cover, TOC, section, sources-section wrapper,
-footer, outlook-badge, design tokens, reset, body / page base).
-
-At emit time, copy the **contents** (not the `<style>` wrapper) of `<style id="shared-template-css">`
-from the shared asset into the parent template's reserved marker region between the CSS-comment
-markers `/* BEGIN shared-template-css ... */` and `/* END shared-template-css */`. For HTML
-scaffolds (head boilerplate, cover, TOC, sources-section wrapper, footer), use the literal markup
-from the matching `<template>` snippet in the shared asset. The parent template no longer
-carries duplicated chrome CSS — those rules ship only in the shared asset.
-
-This skill uses the **`cover-simple`** variant. Skill-specific overrides retained above the
-marker region: **none** for SA (it inherits the canonical `body { font-size: 13px }` and
-`.page { max-width: 900px }` defaults). SA is the lightest of the four parent skills — almost
-all its visual chrome now lives in the shared layer. Any outlook-badge usage in this skill
-(e.g. for sector outlook indicators) must use the canonical pastel variants
-(`stable` / `positive` / `negative` / `review` / `na`) defined by the shared skill — no
-solid-fill or inline-color overrides.
+**Always check the shared template before inventing** any class, token, scaffold, or design element. Do not restyle anything the shared skill owns. At emit time, copy the **contents** of `<style id="shared-template-css">` into the marker region `/* BEGIN shared-template-css ... */` … `/* END shared-template-css */`. Use literal markup from the matching `<template>` snippets for all HTML chrome.
+
+This skill uses the **`cover-simple`** variant with no skill-specific CSS overrides (inherits `body { font-size: 13px }` and `.page { max-width: 900px }`). Outlook-badge usage must use canonical pastel variants (`stable` / `positive` / `negative` / `review` / `na`) — no solid-fill or inline-color overrides.
 
 ## Citations (shared)
 
@@ -99,12 +71,7 @@ Before emitting any `[n]` reference inline or the end-of-document Citations bloc
    canonical CSS (inside `<style id="shared-citations-css">`) and literal HTML markup snippets
    (inside `<template>` tags) for inline references and the end-of-document Citations block.
 
-At emit time, copy the **contents** (not the wrapper) of `<style id="shared-citations-css">`
-from the shared asset into the parent template's reserved marker region, located inside
-`assets/template.html` between the CSS-comment markers
-`/* BEGIN shared-citations-css … */` and `/* END shared-citations-css */`. The parent
-template no longer carries duplicated citation CSS — those rules ship only in the shared
-asset.
+At emit time, copy the **contents** of `<style id="shared-citations-css">` into the marker region `/* BEGIN shared-citations-css … */` … `/* END shared-citations-css */`.
 
 The prefix used for the end-of-document container in this skill is `sa`, so the container id
 is `#sa-sources`. This skill does not use the optional per-section `.section-citations` recap
@@ -321,9 +288,9 @@ Insert **Chart D** immediately after the outlook paragraphs:
 
 ### Pre-computation requirement — mandatory, no exceptions
 
-Before writing any SVG `x`, `y`, `width`, `height`, or `points` attribute that depends on a calculated value (bar lengths, axis tick positions, scale factors, percentage-to-pixel conversions, label offsets, marker positions), compute every value programmatically. **If you do not have a calculator tool, use bash or Python to compute all of the above values before emitting any SVG — do not attempt to calculate them mentally.** Record the computed values, verify them, then substitute the confirmed numbers into the SVG markup.
+Before writing any SVG `x`, `y`, `width`, `height`, or `points` attribute, compute every value via bash or Python — do not calculate mentally. Record and verify computed values, then substitute into the SVG markup.
 
-If source data for any chart is unavailable, emit a `<p>Chart data unavailable</p>` placeholder in its place — never omit it silently.
+If source data for any chart is unavailable, emit a `<p>Chart data unavailable</p>` placeholder — never omit silently.
 
 ---
 
@@ -393,13 +360,8 @@ Each of the following divs takes fully-authored HTML (`<p>`, `<ul><li>`, and
 ## Tips
 
 - Run ALL research searches in a single parallel batch.
-- Moody's data is the foundation — web data supplements. Make this hierarchy clear in the
-  writing by attributing external sources explicitly.
+- Moody's data is the foundation — web data supplements. Attribute external sources explicitly.
 - The executive summary should be written last, after all other sections are synthesized.
-- Financial Performance must include actual numbers, not just directional statements.
-- Porter's Five Forces in Industry Structure should use ratings (High, Moderate, Low) for
-  each force.
-- Sector Outlook must be in paragraph form only — no bullet points.
 - Citations follow the shared citations skill — read
   [skills/shared/citations/SKILL.md](../shared/citations/SKILL.md) before authoring any `[n]`
   reference or the Citations block.
@@ -408,11 +370,7 @@ Each of the following divs takes fully-authored HTML (`<p>`, `<ul><li>`, and
 
 ## Negative number formatting
 
-Unify ALL negative numbers across narrative, tables, and any JSON/HTML string values to use
-the parenthesis-minus convention `(-X)`. Examples: `(-9%)`, `(-1.2pp)`, `(-$3.4B)`,
-`(-250 bps)`. Do NOT use a bare leading minus (`-9%`), en-dash (`–9%`), or accounting
-parentheses without the minus sign (`(9%)`). Apply this consistently to growth rates, margin
-changes, YoY deltas, and any negative monetary values.
+Use `(-X)` for ALL negatives: `(-9%)`, `(-1.2pp)`, `(-$3.4B)`, `(-250 bps)`. Never use bare minus (`-9%`), en-dash (`–9%`), or parentheses without minus (`(9%)`). Apply to growth rates, margin changes, YoY deltas, and negative monetary values.
 
 ---
 
diff --git a/plugins/moody-s/skills/shared/citations/SKILL.md b/plugins/moody-s/skills/shared/citations/SKILL.md
index 90e8446b..ac82c615 100644
--- a/plugins/moody-s/skills/shared/citations/SKILL.md
+++ b/plugins/moody-s/skills/shared/citations/SKILL.md
@@ -4,7 +4,7 @@ description: >
   Canonical rules and HTML/CSS contract for inline `[n]` citation references, end-of-document
   Citations blocks, and optional per-section citation recaps used across Moody's Agentic
   Solutions HTML report skills (earnings-brief, peer-analysis, issuer-brief,
-  sector-brief). Parent skills must read BOTH this `SKILL.md` (rules, numbering, hyperlink
+  sector-brief, etc.). Parent skills must read BOTH this `SKILL.md` (rules, numbering, hyperlink
   behavior, source data shape) AND `assets/template.html` (canonical CSS block + literal HTML
   markup snippets) before emitting citations. The asset file is the single source of truth for
   the visual/markup implementation; this `SKILL.md` is the single source of truth for the
@@ -17,7 +17,7 @@ description: >
 
 This is the **single source of truth** for how citations are authored and rendered across the
 HTML report skills. Parent skills (`earnings-brief`, `peer-analysis`,
-`issuer-brief`, `sector-brief`) defer to this document for inline reference
+`issuer-brief`, `sector-brief`, etc.) defer to this document for inline reference
 markup, the Citations block, numbering, hyperlinking, source data shape, and CSS. Each parent
 skill keeps only its own carve-outs (e.g. "no citations in `.yoy-change` cells", "no citations
 in financial table cells").
@@ -89,7 +89,7 @@ If only one of `source` or `date` is present, render only that field inside the
 
 ## 3. Optional per-section recap (opt-in)
 
-Some skills (currently `earnings-brief`, `peer-analysis`, `issuer-brief`)
+Some skills (currently `earnings-brief`, `peer-analysis`, `issuer-brief`, etc.)
 render a short recap of the citations referenced inside a given section, immediately below
 that section's content. This component is **opt-in**. Parent skills decide whether to use it;
 the canonical CSS in `assets/template.html` ships the styling unconditionally so the option
diff --git a/plugins/moody-s/skills/shared/citations/assets/template.html b/plugins/moody-s/skills/shared/citations/assets/template.html
index f6aae9d8..4d4b8ffd 100644
--- a/plugins/moody-s/skills/shared/citations/assets/template.html
+++ b/plugins/moody-s/skills/shared/citations/assets/template.html
@@ -3,7 +3,7 @@
 
   This file is the SINGLE SOURCE OF TRUTH for the visual contract of citations
   across the HTML report skills (earnings-brief, peer-analysis,
-  issuer-brief, sector-brief).
+  issuer-brief, sector-brief, etc.).
 
   Parent skills MUST:
     1. Read this file alongside skills/shared/citations/SKILL.md before emitting
diff --git a/plugins/moody-s/skills/shared/template/SKILL.md b/plugins/moody-s/skills/shared/template/SKILL.md
index edeb50fe..908394de 100644
--- a/plugins/moody-s/skills/shared/template/SKILL.md
+++ b/plugins/moody-s/skills/shared/template/SKILL.md
@@ -1,24 +1,23 @@
 ---
 name: template
 description: >
-  Canonical rules and HTML/CSS contract for page chrome (head boilerplate, cover, table of
-  contents, section block, sources-section wrapper, footer, outlook-badge, and design tokens)
-  used across Moody's Agentic Solutions HTML report skills (earnings-call-summary, peer-analysis,
-  public-information-book, sector-analysis). Sibling of `skills/shared/citations/` using the
-  same inlining pattern. Parent skills must read BOTH this `SKILL.md` (owned classes/snippets,
-  per-skill overrides, outlook-badge usage) AND `assets/template.html` (canonical CSS + HTML
-  markup snippets) before emitting any HTML report. Asset file is the single source of truth
-  for visual/markup; this `SKILL.md` is the single source of truth for authoring rules. Triggers
-  when the user asks about the template, page chrome, cover, table of contents, footer, layout,
-  design tokens, color palette, fonts, outlook badges, or anything affecting the visual
-  scaffolding of an HTML report skill.
+  Canonical rules and HTML/CSS contract for the page chrome (head boilerplate, cover, table of
+  contents, section block, sources-section wrapper, footer, outlook-badge, design tokens) shared
+  across Moody's Agentic Solutions HTML report skills (earnings-call-summary, peer-analysis,
+  public-information-book, sector-analysis, etc.). Sibling of `skills/shared/citations/` using the same
+  inlining pattern. Parent skills must read BOTH this `SKILL.md` (authoring rules: shared-layer
+  classes, allowed per-skill overrides, outlook-badge usage) AND `assets/template.html` (canonical
+  CSS + literal HTML snippets) before emitting any report. The asset is the source of truth for
+  markup; this file is the source of truth for rules. Triggers on questions about the template,
+  page chrome, cover, table of contents, footer, layout, design tokens, color palette, fonts,
+  outlook badges, or any visual scaffolding of an HTML report skill.
 ---
 
 # Template Skill (shared)
 
 This is the **single source of truth** for the visual chrome of every HTML report produced by
-the parent skills `earnings-brief`, `peer-analysis`, `issuer-brief`, and
-`sector-brief`. It owns:
+the parent skills `earnings-call-summary`, `peer-analysis`, `public-information-book`, and
+`sector-analysis` and others. It owns:
 
 - Design tokens (`--navy`, `--accent`, `--gray-*` / `--g*`, `--text`, `--positive`, `--negative`,
   `--neutral`).
@@ -89,8 +88,8 @@ Two variants. Each parent skill uses exactly one:
 
 | variant       | used by                                  | shape                                            |
 |---------------|------------------------------------------|--------------------------------------------------|
-| `cover-simple`| `issuer-brief`, `sector-brief` | 1-column navy cover, top "logo + title" / bottom `.cover-meta`. |
-| `cover-multi` | `earnings-brief`, `peer-analysis` | 2-region grid: `.cover-top` with navy `.cover-left` + optional `.cover-strip` image + `.cover-bar-top`; `.cover-bottom` is an optional landscape band. |
+| `cover-simple`| `public-information-book`, `sector-analysis` | 1-column navy cover, top "logo + title" / bottom `.cover-meta`. |
+| `cover-multi` | `earnings-call-summary`, `peer-analysis` | 2-region grid: `.cover-top` with navy `.cover-left` + optional `.cover-strip` image + `.cover-bar-top`; `.cover-bottom` is an optional landscape band. |
 
 Cover-multi rules:
 
@@ -162,10 +161,12 @@ the marker region (and only those — anything else inside the canonical block i
 
 | skill                       | allowed local overrides                                                |
 |-----------------------------|------------------------------------------------------------------------|
-| `earnings-brief`            | none. Cover variant: `cover-multi`.                                    |
+| `earnings-call-summary`     | none. Cover variant: `cover-multi`.                                    |
 | `peer-analysis`             | `.page { max-width: 1050px }`. Cover variant: `cover-multi`.           |
-| `issuer-brief`              | `body { font-size: 12.5px }`, `.page { max-width: 920px }`. Cover variant: `cover-simple`. |
-| `sector-brief`              | none. Cover variant: `cover-simple`.                                   |
+| `public-information-book`   | `body { font-size: 12.5px }`, `.page { max-width: 920px }`. Cover variant: `cover-simple`. |
+| `sector-analysis`           | none. Cover variant: `cover-simple`.                                   |
+| `rating-analysis`           | Skill-managed template (Python builder `scripts/build_html.py`). Does not inline shared-template-css or shared-citations-css. Chart.js loaded from CDN. Cover variant: n/a.|
+
 
 Skill-specific CSS that the shared layer does not own (and which therefore stays in each
 parent template indefinitely):
diff --git a/plugins/moody-s/skills/shared/template/assets/template.html b/plugins/moody-s/skills/shared/template/assets/template.html
index d8c35cbb..af1bbb7c 100644
--- a/plugins/moody-s/skills/shared/template/assets/template.html
+++ b/plugins/moody-s/skills/shared/template/assets/template.html
@@ -4,7 +4,7 @@
   This file is the SINGLE SOURCE OF TRUTH for the visual contract of the page
   chrome (head boilerplate, cover, TOC, section, sources wrapper, footer,
   outlook-badge) across the HTML report skills (earnings-brief,
-  peer-analysis, issuer-brief, sector-brief). It is the sibling
+  peer-analysis, issuer-brief, sector-brief, etc.). It is the sibling
   of skills/shared/citations/assets/template.html and uses the exact same
   inlining contract.