docs: clarify Cowork vs Claude Code settings file location#124
Merged
tobinsouth merged 1 commit intomainfrom Mar 18, 2026
Merged
docs: clarify Cowork vs Claude Code settings file location#124tobinsouth merged 1 commit intomainfrom
tobinsouth merged 1 commit intomainfrom
Conversation
The READMEs instructed users to place settings files in 'your project's .claude/ directory' — but Cowork has no project concept. Its .claude/ is ephemeral per-session app storage in ~/Library/Application Support/ that users can't practically access. Cowork users should instead place settings files in any folder shared with Cowork via the folder picker. The skill implementations already search freely with no hardcoded path, so this is purely a docs fix. Reported via Reddit: https://www.reddit.com/r/ClaudeCowork/comments/1rh875p/
tobinsouth
approved these changes
Mar 18, 2026
cbellbell-spin
added a commit
to cbellbell-spin/knowledge-work-plugins
that referenced
this pull request
Apr 17, 2026
* Add pdf-research plugin with local MCP server First knowledge-work plugin to use a local stdio MCP server instead of a remote HTTP connector. The pdf-server runs locally via npx and provides interactive PDF viewing for academic papers and local files. Tested end-to-end in Claude Code: plugin validation, MCP server connection, tool availability (list_pdfs, display_pdf), and command/skill registration all confirmed working. Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com> * chore: update plugin-authoring skills to use skills/ format The CoWork UI now presents commands and skills as a single 'Skills' concept. Update the create-cowork-plugin and cowork-plugin-customizer skills to scaffold skills/*/SKILL.md instead of commands/*.md. The legacy commands/ format still works. * docs: fix version frontmatter placement — must be under metadata per spec Addresses review feedback. Per https://agentskills.io/specification#frontmatter, version is not a top-level frontmatter field — it goes under the arbitrary metadata map. Top-level version gets rejected by validators. * Migrate commands to skills across all plugins - Convert all commands/*.md files to skills/*/SKILL.md format - Merge 36 reference-skill pairs into their invokable counterparts - Mark reference-only skills as user-invocable: false - Expand descriptions for migrated skills with usage scenarios - Keep partner-built/ commands as-is - Remove bio-research from marketplace.json * Bump plugin versions after commands-to-skills migration Bump minor version across all 14 plugins to reflect the migration from commands/ to skills/ format shipped in the previous merge. * Update disclaimer with jurisdictional disclaimer (anthropics#35) Expanded disclaimer to include jurisdictional customization requirements for the plugin. * Add manual plugin validation fallback when CLI validator is unavailable (anthropics#63) When running inside Cowork, `claude plugin validate` may not be available. This adds a manual verification checklist covering plugin.json structure, naming conventions, component directories, and expected file formats. Co-authored-by: Claude Opus 4.6 <noreply@anthropic.com> * fix: replace 5 bare excepts with except Exception in bio-research scripts (anthropics#82) Bare `except:` clauses catch `KeyboardInterrupt` and `SystemExit`, which can mask real errors and make debugging harder. Replace with `except Exception:` to preserve the intended error-suppression behavior while allowing system-level exceptions to propagate normally. Files changed: - bio-research/skills/nextflow-development/scripts/check_environment.py (4 sites) - bio-research/skills/nextflow-development/scripts/detect_data_type.py (1 site) Co-authored-by: haosenwang1018 <haosenwang1018@users.noreply.github.com> * feat(product-management): add brainstorming skill and /brainstorm command (anthropics#93) * feat(product-management): add brainstorming skill and /brainstorm command The product-management plugin covers structured PM outputs (specs, roadmaps, stakeholder updates) but lacks support for the divergent thinking that precedes them. This adds a brainstorming skill and command for exploring problem spaces, generating ideas, and stress-testing product thinking. New files: - skills/product-brainstorming/SKILL.md — brainstorming modes (problem exploration, solution ideation, assumption testing, strategy exploration), PM frameworks (HMW, JTBD, First Principles, SCAMPER, Opportunity Solution Trees, Reverse Brainstorming), session structure, thinking partner behaviors, and anti-patterns - commands/brainstorm.md — conversational brainstorming command with tool-augmented context gathering and structured session flow Updated: - README.md — added brainstorming to What It Does, Commands, Skills tables, and Example Workflows Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com> * Add OODA Loop framework to product-brainstorming skill Adds Observe-Orient-Decide-Act as a brainstorming framework for competitive and time-pressured product decisions. Focuses on decision tempo and when to use OODA to break analysis paralysis in sessions. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com> --------- Co-authored-by: Claude Opus 4.6 <noreply@anthropic.com> * docs: clarify Cowork vs Claude Code settings file location (anthropics#124) The READMEs instructed users to place settings files in 'your project's .claude/ directory' — but Cowork has no project concept. Its .claude/ is ephemeral per-session app storage in ~/Library/Application Support/ that users can't practically access. Cowork users should instead place settings files in any folder shared with Cowork via the folder picker. The skill implementations already search freely with no hardcoded path, so this is purely a docs fix. Reported via Reddit: https://www.reddit.com/r/ClaudeCowork/comments/1rh875p/ * Add Supermetrics as a marketing analytics connector to the marketing plugin (anthropics#123) * Add Definite as a data warehouse connector (anthropics#31) Definite is an all-in-one analytics platform (DuckDB warehouse, semantic modeling, BI, and AI) with an HTTP MCP server. Adds it to the data plugin alongside existing warehouse connectors. Co-authored-by: Claude Opus 4.6 <noreply@anthropic.com> * Add Amplitude EU endpoint option (anthropics#7) Add amplitude-eu MCP server configuration for EU users alongside the existing US endpoint in data, marketing, and product-management. Co-authored-by: Claude Opus 4.5 <noreply@anthropic.com> * Plugin batch: 12 new plugins (intercom, cockroachdb, prisma, fastly, cloudinary, nimble, brightdata, searchfit-seo, atlan, ai-firstify, product-tracking, postiz) (anthropics#117) * Add planetscale to marketplace * Add adspirer-ads-agent to marketplace * Add sanity-plugin to marketplace * Add zoominfo to marketplace * Add mintlify to marketplace * add(plugin-json): daloopa * Add zapier to marketplace * Add intercom to marketplace * Add cockroachdb to marketplace * Add prisma to marketplace * Add fastly-agent-toolkit to marketplace * Add cloudinary to marketplace * Add nimble to marketplace * Add brightdata-plugin to marketplace * Add searchfit-seo to marketplace * Add atlan to marketplace * Add ai-firstify to marketplace * Add product-tracking-skills to marketplace * Add postiz to marketplace * Add Figma plugin to marketplace --------- Co-authored-by: Claude <noreply@anthropic.com> * Rescope pdf-research → pdf: expose annotation, form-filling, signing The underlying @modelcontextprotocol/server-pdf gained an `interact` tool with 10 annotation types (highlight, underline, strikethrough, note, rectangle, circle, line, freetext, stamp, image), form filling, auto-highlight-text, page extraction, and annotated-PDF download. This reshapes the plugin around interactive workflows: - Rename plugin: pdf-research → pdf - Rename skill: pdf-reading → pdf (unified, covers all capabilities) - Commands: drop /summarize, rename /read → /open, add /annotate, /fill-form, /sign - SKILL.md: document interact tool, all annotation types, coordinate system, interactive workflows (AI-driven annotate, form filling, visual signing) - Positioning: explicitly steer AWAY from viewer for pure ingestion (use native Read instead) — the viewer's value is user interactivity - /sign disclaimer: visual signature image, not certified digital sig Out of scope (documented): summarization, cert signing, PDF generation. * Address review: fix accuracy, conventions, emphasize visual form-filling Accuracy: - list_pdfs: 'remote origins' → 'local directories' (server has no remote allowlist, any HTTPS works) - get_screenshot: 'PNG' → 'image' (server returns JPEG) - Supported sources: clarify only arXiv auto-converts /abs/→PDF; others need direct PDF URLs Conventions (matching legal/, productivity/ patterns): - Add argument-hint frontmatter to all 4 commands - Add CONNECTORS.md callout to all commands Positioning — form filling: - Emphasize this is VISUAL form filling (vs programmatic tools) - Document the unnamed-field workflow: screenshot → match bounding boxes to visual labels → fill by name. Many real PDFs have fields named 'Text1', 'Field_7' with labels only on the rendered page. - User gets live feedback and can edit directly in viewer * Rename pdf → pdf-viewer, skill → view-pdf; refine trigger description Plugin name: pdf → pdf-viewer - 'Viewer' makes the interactive/visual modality self-evident - Naturally explains why NOT to use for summarization - Matches enterprise-search precedent (capability-named, not job-function) Skill name: pdf → view-pdf - Verb-noun matches repo convention (review-contract, triage-nda) Trigger description (SKILL.md frontmatter): - Hybrid style: situation framing + verb list for keyword matching - Includes 'open, show, view' so bare 'open this PDF' triggers the skill — SKILL body then disambiguates viewer vs Read - Keeps 'Not for summarization (use Read)' guard Marketplace/plugin.json description: - Marketing-oriented (benefits-first) to match other plugins' style - 'View, annotate, sign... Mark up contracts, fill forms with visual feedback, stamp approvals, place signatures — download annotated copy' Updated all /pdf:* cross-references → /pdf-viewer:* * Add -y flag to npx to avoid first-install prompt hanging MCP connection Without -y, npx prompts 'Ok to proceed? (y)' on first install of a package, which hangs the stdio MCP connection (no TTY to answer). Keeping version unpinned (auto-latest) — currently pulls 1.3.1 which has the interact/annotation capabilities. Tested end-to-end via 'claude --plugin-dir pdf-viewer/'. * Restore bio-research to marketplace.json (anthropics#165) The bio-research entry was removed in 2d6f7e2 (PR anthropics#115) without a stated reason, but the plugin directory was migrated and continues to be maintained. This restores the original entry pending confirmation that the removal was intentional. * add(plugin-json): zoom-plugin external source (anthropics#175) * Add Zoom partner plugin (zoom-plugin) (anthropics#174) Official Zoom plugin with 30 skills covering the Zoom Developer Platform: REST APIs, Meeting SDK, Video SDK, webhooks, OAuth, bots, and MCP workflows. Bundles 3 remote MCP connectors (meetings, docs, whiteboard) hosted at zoom.us. Zoom MCP connector passed internal review (Anthony Bibbs, 2026-04-07). Co-authored-by: Claude Opus 4.6 (1M context) <noreply@anthropic.com> * Revert "add(plugin-json): zoom-plugin external source (anthropics#175)" (anthropics#176) This reverts commit 3f12086. * Add missing resume-tailor skill to plugin bundle (v0.4.3) The resume-tailor skill was included in the v0.4.2 CHANGELOG but was not added to the plugin archive. This change adds it to the bundle. Note: Years-of-experience and graduation-date integrity rules are NOT included — available via post-install customization. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com> --------- Co-authored-by: Bryan Thompson <bthompson@anthropic.com> Co-authored-by: Claude Opus 4.6 (1M context) <noreply@anthropic.com> Co-authored-by: Henry Shi <henrys@anthropic.com> Co-authored-by: Henry Shi <henrythe9th@gmail.com> Co-authored-by: Matt Piccolella <mattpic@anthropic.com> Co-authored-by: Adil Anwar <107459552+TheGreatAdil@users.noreply.github.com> Co-authored-by: hichana <8943047+hichana@users.noreply.github.com> Co-authored-by: Sense_wang <167664334+haosenwang1018@users.noreply.github.com> Co-authored-by: haosenwang1018 <haosenwang1018@users.noreply.github.com> Co-authored-by: Naren Katakam <naren.katakam@gmail.com> Co-authored-by: vinay <vinayprajagopal@gmail.com> Co-authored-by: bartschneider <84577919+bartschneider@users.noreply.github.com> Co-authored-by: Mike Ritchie <96531430+mike-luabase@users.noreply.github.com> Co-authored-by: Simone Basso <smnbss@hotmail.com> Co-authored-by: Tobin South <tobin.south@gmail.com> Co-authored-by: Olivier Chafik <ochafik@anthropic.com> Co-authored-by: Jason Morrison <jason.p.morrison@gmail.com> Co-authored-by: Chris Bell <chris@chrisjbell.dev>
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Problem
All three plugin READMEs instruct users to place settings files in
.claude/directories — but these are Cowork-first plugins (each README says "primarily designed for Cowork" on line 3), and that instruction doesn't work in Cowork.Why
.claude/doesn't work in Cowork: Cowork has no project concept. Its.claude/directory is ephemeral, per-session, app-managed storage in~/Library/Application Support/Claude/.../{sessionId}/.claude/— not a location users can practically place files in. The agent's cwd is an empty/sessions/{name}home.What Cowork users should do instead: Save the settings file in any folder shared with Cowork via the folder picker (mounted at
$HOME/mnt/{name}in the VM). The skill implementations already search freely with no hardcoded path, so they'll find it.Reported via Reddit: https://www.reddit.com/r/ClaudeCowork/comments/1rh875p/
Fix
Replaced the single-surface path instruction with a Cowork / Claude Code split:
.claude/directory (or the plugin-specific path, depending on the README).This is docs-only — no code changes needed. The skill implementations (
SKILL.md) already search freely.Files changed
legal/README.md— "Configure your playbook" section (legal.local.md)human-resources/README.md— "Settings" section (settings.local.json)sales/README.md— "Settings" section (settings.local.json)Conflicts note
PRs #119 (install command syntax) and #110 (legal jurisdiction disclaimer) also touch
legal/README.md, but in different sections — no textual overlap expected.