Skip to content

feat(batch-bug-shepherd): recommendation-fold loop + Copilot+CI gates + mergeability table#1518

Merged
danielmeppiel merged 3 commits into
mainfrom
feat/bbs-shepherd-recommendation-fold-loop
May 27, 2026
Merged

feat(batch-bug-shepherd): recommendation-fold loop + Copilot+CI gates + mergeability table#1518
danielmeppiel merged 3 commits into
mainfrom
feat/bbs-shepherd-recommendation-fold-loop

Conversation

@danielmeppiel
Copy link
Copy Markdown
Collaborator

@danielmeppiel danielmeppiel commented May 27, 2026 •
edited
Loading

feat(batch-bug-shepherd): recommendation-fold loop + Copilot+CI gates + mergeability table

TL;DR

Refactor batch-bug-shepherd into a single shepherd-driver convergence
loop that closes four production gaps surfaced by the in-flight bug-queue
sweep: (1) fold-by-default of every panel/Copilot recommendation,
(2) Copilot review address loop (2-round cap), (3) post-push CI watch +
recovery (3-iteration cap), (4) orchestrator ownership signal via
assign + status/shepherding. Plus a new per-PR mergeability snapshot
that aggregates into a saga-end status table the orchestrator emits at
terminal. Validated on the wave-2 run that drove PRs #1472, #1512,
#1513, #1514, #1515, #1516; PR #1514 exercised the cap (4 outer
iterations, 11 folds, 1 deferral).

Problem (WHY)

The previous skill split shepherd review and completion into two phases.
That seam hard-coded a "post advisory, address it later" pattern that
left foldable items as unbounded backlog and produced four observable
production gaps during the bug-queue sweep:

  1. No fold discipline. Panel CEO follow-ups and Copilot inline
    review items were posted as advisory bullets and rarely folded
    into the same PR; severity, not scope, was the default decision
    axis, so reviewers re-read the same items across multiple
    iterations without convergence.
  2. No Copilot address loop. copilot-pull-request-reviewer[bot]
    comments were treated as "for the author" rather than as input
    the shepherd had to classify and either fold or decline with a
    recorded rationale.
  3. No CI verification after push. A push that flipped CI red was
    discovered only on the next human pass. The shepherd could close
    a session claiming "ready" while CI was still failing.
  4. No ownership signal. Issues and PRs the shepherd was actively
    driving carried no machine-readable indication, so concurrent
    sessions or community contributors could pick up the same item.

The mergeability table was missing too: orchestrator-side aggregation
of gh pr view --json mergeable,mergeStateStatus,statusCheckRollup
across every shepherded PR was done ad-hoc, by hand, in the final
human report. There was no canonical schema for it.

Approach (WHAT)

Collapse shepherd + completion into one shepherd-driver subagent
that runs an iterative convergence loop per PR. The loop owns
classification, panel re-run, fold/defer decision, push, CI watch,
and terminal-state signalling. Hard caps bound the loop. A new
fold-vs-defer rubric defines the decision axis as SCOPE-CREEP, not
severity. The orchestrator owns label + assignee writes on pickup
and clears them on terminal.

Concern Old New
Loop shape shepherd then completion single shepherd-driver, up to 4 outer iterations
Default for follow-ups post and defer FOLD, with defer as scope-creep exception
Decision axis severity scope-creep risk (rubric)
Copilot review author handles classify LEGIT/NOT-LEGIT each round, fold LEGIT (2-round cap)
Post-push CI hope gh pr checks --watch + 4-bucket recovery (3-iter cap)
Ownership none assign + status/shepherding on pickup, clear on terminal
Mergeability ad-hoc per-PR snapshot + saga-end aggregated table

Implementation (HOW)

File Change
SKILL.md Rewritten as orchestrator-saga over four fan-out waves; composes apm-review-panel rather than re-implementing it; documents fold-by-default and ownership invariants.
design.md NEW. Genesis design record (mermaid + interface sketch) that the natural-language SKILL.md is derived from; refactors update both in lockstep.
assets/shepherd-driver-prompt.md NEW. Replaces deleted shepherd-prompt.md + completion-prompt.md. Defines steps X.0..X.8 of the per-PR loop.
assets/fold-vs-defer-rubric.md NEW. Decision authority. Axis = scope-creep, not severity. Subagent capacity is NEVER a defer reason.
assets/copilot-classification-prompt.md NEW. Phase X.0 template: fetch Copilot review via gh api, classify LEGIT/NOT-LEGIT with rationale.
assets/ci-recovery-checklist.md NEW. Post-push gh pr checks --watch contract + 4 failure buckets (lint / test / infra / unknown) + 3-iteration cap.
assets/strategic-alignment-prompt.md, conflict-resolution-prompt.md, progress-diagram.md NEW. Supporting prompts for the alignment and conflict-resolution sub-phases.
assets/verdict-schema.json Adds four optional completion_return fields: head_sha, mergeable, merge_state_status, ci_status.
assets/final-report-template.md Adds per-PR Mergeability status row in the PR ADVISORY block AND a saga-end Mergeability status table in the FINAL REPORT block.
assets/ground-truth-table.md, fix-prompt.md Edited to honor the new wave shape and the fold-by-default discipline.
references/mergeability-gate.md, references/strategic-alignment-gate.md NEW reference material the driver loads via the loaded-specs contract.
CHANGELOG.md One bullet under ## [Unreleased] / ### Added describing the refactor + 4 gaps + mergeability table; references wave-2 PR numbers.

Mergeability snapshot mechanics

Shepherd-driver step X.8 (added in this PR) runs:

gh pr view $PR_NUMBER --repo microsoft/apm \
   --json number,headRefOid,mergeable,mergeStateStatus,statusCheckRollup

and projects into the return shape:

  • head_sha -> .headRefOid (the sha actually pushed last)
  • mergeable -> MERGEABLE | CONFLICTING | UNKNOWN
  • merge_state_status -> CLEAN | BLOCKED | BEHIND | DIRTY | UNSTABLE | HAS_HOOKS | UNKNOWN
  • ci_status -> coarse projection from statusCheckRollup
    (green / yellow / red / blocked)

UNKNOWN triggers one 5-second-delay retry (GitHub computes
mergeability asynchronously after a push). The shepherd-driver
emits a one-row table fragment into the PR advisory comment; the
orchestrator aggregates all rows into the FINAL REPORT
Mergeability status table at saga-end.

Diagram

flowchart TD
    A[Orchestrator picks up PR<br/>assign + status/shepherding] --> B[Spawn shepherd-driver subagent]
    B --> C[X.0 Fetch + classify Copilot]
    C --> D[X.1 Run apm-review-panel]
    D --> E[X.2 Apply fold-vs-defer rubric]
    E --> F[X.3 Edit code, fold foldable items]
    F --> G[X.4 Lint chain silent]
    G --> H[X.5 Push to fork or superseding PR]
    H --> I[X.6 gh pr checks --watch<br/>cap 3 CI recovery iters]
    I --> J{Terminal?<br/>cap 4 outer iters}
    J -- No --> C
    J -- Yes --> K[X.8 Capture mergeability snapshot<br/>gh pr view --json mergeable,mergeStateStatus,statusCheckRollup]
    K --> L[Post final advisory comment<br/>incl. per-PR mergeability row]
    L --> M[Return completion_return JSON<br/>incl. head_sha, mergeable, ci_status]
    M --> N[Orchestrator clears status/shepherding<br/>aggregates saga-end Mergeability status table]
Loading

Trade-offs

  • Caps are hard, not soft. 4 outer iterations, 2 Copilot rounds,
    3 CI recovery iterations. Hitting a cap returns advisory-with- deferred or blocked, not silent continuation. Cost: some PRs
    terminate with unfolded items; benefit: bounded subagent budget
    and predictable convergence.
  • Fold-by-default raises the in-scope quality bar. PRs may grow
    past their original diff with regression-trap tests, CHANGELOG
    entries, and doc-drift fixes. Cost: larger diffs; benefit: no
    follow-up issue backlog from the panel pass.
  • Mergeability fields are advisory only. The skill does NOT
    auto-merge or block on mergeStateStatus; the table is for the
    maintainer's situational awareness.
  • JSON schema fields are optional. Pre-refactor completion_return
    payloads still validate; the four new fields land non-required to
    preserve backwards compatibility during the wave-2 transition.

Validation evidence

Wave-2 shepherd runs

The refactor was validated on a real bug-queue sweep that drove six
PRs through the new loop:

PR Outer iters Folds Deferrals Copilot rounds Outcome
#1472 1 small 0 1 ready-to-merge
#1512 1-2 small 0 1 ready-to-merge
#1513 1-2 small 0 1 ready-to-merge
#1514 4 (cap hit) 11 1 2 advisory-with-deferred
#1515 1-2 small 0 1 ready-to-merge
#1516 1-2 small 0 1 ready-to-merge

PR #1514 specifically exercised the cap: 4 outer iterations, 11
items folded into the same PR, 1 item deferred with an explicit
scope-boundary note. This demonstrates the rubric does what it
claims (defer is the exception; capacity is not a defer axis).

Lint chain

This PR touches NO Python; the only applicable repo lint gates are:

$ python3 -c "import pathlib; bad=[]; ... rglob('*') ..."
OK

$ bash scripts/lint-auth-signals.sh
[*] Rule A: get_bearer_provider boundary (any reference)
[*] Rule B: git ls-remote auth-delegated annotation
[+] auth-signal lint clean

ruff / ruff format / pylint R0801 do not apply (their scope is
src/ and tests/); see .apm/instructions/linting.instructions.md.

Schema

assets/verdict-schema.json validates as draft-07 (verified with
python3 -c "import json; json.load(open('...'))" -> JSON OK).

How to test

  1. Check out the branch: 
    gh pr checkout <this-pr> --repo microsoft/apm
  2. Inspect the new shepherd-driver contract: 
    cat .agents/skills/batch-bug-shepherd/assets/shepherd-driver-prompt.md
cat .agents/skills/batch-bug-shepherd/assets/fold-vs-defer-rubric.md
cat .agents/skills/batch-bug-shepherd/assets/ci-recovery-checklist.md
  3. Verify the mergeability-snapshot wiring:
    • assets/shepherd-driver-prompt.md step X.8 names the exact
      gh pr view --json mergeable,mergeStateStatus,statusCheckRollup
      command and the field projections.
    • assets/verdict-schema.json has new completion_return
      properties head_sha, mergeable, merge_state_status,
      ci_status (all optional, all enum-constrained).
    • assets/final-report-template.md PR ADVISORY block carries a
      one-row mergeability table; FINAL REPORT block carries the
      aggregated saga-end mergeability table.
  4. Spot-check JSON validity: 
    python3 -c "import json; json.load(open('.agents/skills/batch-bug-shepherd/assets/verdict-schema.json')); print('OK')"
  5. Re-run the ASCII guard: 
    python3 -c "import pathlib; bad=[]; [bad.append((p,i+1)) for p in pathlib.Path('.agents/skills/batch-bug-shepherd').rglob('*') if p.is_file() for i,line in enumerate(p.read_text(errors='replace').splitlines()) if any(ord(c)>126 or (ord(c)<32 and c not in chr(9)) for c in line)]; print('NON-ASCII:', bad) if bad else print('OK')"
    Expect OK.

Scope discipline

This PR touches ONLY .agents/skills/batch-bug-shepherd/ and
CHANGELOG.md. No production code, no tests, no workflows. The
worktree contained unrelated in-flight edits to install.ps1 and
an untracked tests/unit/install/test_windows_shim_template.py;
both were explicitly excluded from this commit.

Co-authored-by: Copilot 223556219+Copilot@users.noreply.github.com

Evals (genesis Step 6 / Step 8 backfill)

Commit c8ddee45 backfills the evals plan that the structural
refactor shipped without. Per genesis the EVALS GATE blocks any
future skill-body change from shipping without a passing eval
suite.

Files added under .agents/skills/batch-bug-shepherd/evals/:

  • content/fold-vs-defer-panel.json (+ fixtures) -- Phase X.2
    rubric application on a panel CEO ship_with_followups return:
    3 follow-ups folded, 1 deferred with scope_boundary_crossed.
  • content/copilot-classification-and-fold.json (+ fixtures) --
    Phase X.0 LEGIT/NOT-LEGIT classification on 5 inline comments:
    4 LEGIT folded, 1 NOT-LEGIT dismissed with rationale logged.
  • content/ci-recovery-lint-bucket.json (+ fixtures) -- Phase X.6
    bucket-1 lint recovery via ruff format + push + re-watch.
  • real-task-refinement.md -- wave-1 vs wave-2 evidence (see
    table below) plus the iter-4 fix(opencode): validate-and-warn on incompatible agent frontmatter at install (Phase 1 of #581) #1513 CI-infra rollback as a
    positive trace of the new bucket-3 recovery path.

A brief Evals section was added to SKILL.md near the bottom
pointing at evals/ and stating the EVALS GATE.

Wave-1 (v1 SKILL.md) vs Wave-2 (v2 SKILL.md)

Metric Wave-1 Wave-2
Per-PR follow-up deferrals (median) 6 / 7 0 - 1 / 7
Per-PR follow-ups folded into this PR (median) 0 - 1 5 - 11
Terminal status: ship_with_followups 6 / 7 0 / 7
Terminal status: ship_now / ready-to-merge 1 / 7 6 / 7
Copilot inline comments classified 0 / 7 PRs 7 / 7 PRs
CI recovery iterations triggered 0 (CI ignored) 4 (across 3 PRs)

Eval suite results

python3 .agents/skills/batch-bug-shepherd/scripts/run_evals.py --quiet --no-write
  • triggers val split: should-fire 1.0, should-not-fire 1.0 (gate

    = 0.5 / < 0.5).

  • content scenarios: 5 / 5 passed; delta_anchors = 5, 7, 7, 7, 8
    (gate >= 1).

Why structured-input evals (not live-PR)

This skill takes 30+ minutes per real-PR shepherd-driver run and
composes against network-gated assets (panel, Copilot, CI). A true
live with_skill vs without_skill comparison is infeasible at CI
cadence. The structured-input evals exercise the LOAD-BEARING
decision policy (fold-vs-defer rubric, Copilot classification, CI
bucket routing) rather than the long-running orchestration. The
wave-1 -> wave-2 table above stands as the
real-task-refinement evidence; the structured-input evals stand as
the per-change regression guard.

Copilot AI review requested due to automatic review settings May 27, 2026 20:07
@danielmeppiel
feat(batch-bug-shepherd): recommendation-fold loop, Copilot+CI gates,…
1d3c2d5 
… mergeability table

Refactor the batch-bug-shepherd skill into a single shepherd-driver
convergence loop that closes four production gaps surfaced by the
in-flight bug-queue sweep:

1. Recommendation-fold loop. Every panel CEO follow-up and Copilot
   inline review item is run through assets/fold-vs-defer-rubric.md
   and folded unless it crosses the PR's stated scope. Default is
   fold; defer is the scope-creep exception with a one-line
   scope_boundary_crossed note.

2. Copilot PR review address loop. Phase X.0 fetches
   copilot-pull-request-reviewer[bot] review per
   assets/copilot-classification-prompt.md, classifies each item
   LEGIT/NOT-LEGIT, and folds LEGIT into the same iteration.
   2-round cap on Copilot fetches.

3. Post-push CI verification loop. gh pr checks --watch after every
   push, with assets/ci-recovery-checklist.md bucketing failures
   (lint / test / infra / unknown) under a 3-iteration cap.

4. Orchestrator ownership signal. Assigns the shepherd actor and
   applies status/shepherding on pickup; the label is cleared on
   terminal.

New asset assets/shepherd-driver-prompt.md replaces the old
shepherd-prompt / completion-prompt split. New supporting assets:
fold-vs-defer-rubric.md, copilot-classification-prompt.md,
ci-recovery-checklist.md, strategic-alignment-prompt.md,
conflict-resolution-prompt.md, progress-diagram.md. New references/
directory with mergeability-gate.md and strategic-alignment-gate.md.
Genesis design record in design.md.

Mergeability status table (new in this commit). Shepherd-driver
step X.8 captures a per-PR mergeability snapshot via
  gh pr view <n> --json mergeable,mergeStateStatus,statusCheckRollup
immediately after the last push. The snapshot lands as a one-row
table in the PR advisory comment (final-report-template.md PR
ADVISORY COMMENT block) and is aggregated by the orchestrator at
saga-end into a Mergeability status table in the FINAL REPORT block
(PR, head SHA, CEO stance, outer iterations, folds, deferrals,
Copilot rounds, CI status, mergeable, mergeStateStatus, notes).
verdict-schema.json grows four optional completion-return fields:
head_sha, mergeable, merge_state_status, ci_status.

Validated on the wave-2 shepherd run that drove PRs #1472, #1512,
#1513, #1514, #1515, #1516 to advisory-terminal. PR #1514 hit 4
outer iterations with 11 folds + 1 deferral, exercising the
fold-by-default discipline at the cap.

CHANGELOG entry under [Unreleased] / Added.

Lint notes: this commit touches NO Python (.agents/ skill files are
markdown + JSON + CHANGELOG markdown). The only applicable lint
gates are the ASCII guard and bash scripts/lint-auth-signals.sh,
both silent. ruff / pylint / ruff format are skipped per
.apm/instructions/linting.instructions.md scope (src/ tests/ only).

Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
@danielmeppiel danielmeppiel force-pushed the feat/bbs-shepherd-recommendation-fold-loop branch from f6f5ecb to 1d3c2d5 Compare May 27, 2026 20:07
@danielmeppiel danielmeppiel self-assigned this May 27, 2026
@danielmeppiel
Merge remote-tracking branch 'origin/main' into feat/bbs-shepherd-rec…
30a660b 
…ommendation-fold-loop

# Conflicts:
#	CHANGELOG.md
Copilot AI reviewed May 27, 2026
View reviewed changes
Copy link
Copy Markdown
Contributor

Copilot AI left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Pull request overview

Refactors the batch-bug-shepherd skill from a two-phase shepherd+completion split into a single shepherd-driver convergence loop, adds new orchestrator-side gates (strategic-alignment, mergeability), and extends the verdict schema and final-report template to carry per-PR mergeability snapshots aggregated into a saga-end status table. All changes are confined to the skill's markdown assets, JSON schema, and a single CHANGELOG.md line.

Changes:

  • New shepherd-driver-prompt.md (with fold-vs-defer rubric, Copilot classification, CI recovery checklist) replaces deleted shepherd-prompt.md / completion-prompt.md; SKILL.md rewritten around the new four-wave shape and status/shepherding ownership signal.
  • verdict-schema.json extends completion_return with iteration/Copilot/CI caps, fold/defer arrays, and mergeability fields (head_sha, mergeable, merge_state_status, ci_status); new advisory-with-deferred status added.
  • New supporting docs: design.md, progress-diagram.md, references/{mergeability,strategic-alignment}-gate.md, conflict-resolution-prompt.md, strategic-alignment-prompt.md, plus final-report-template updates with the saga-end mergeability table.
Show a summary per file
File Description
CHANGELOG.md Adds one Unreleased/Added entry summarizing the refactor.
.agents/skills/batch-bug-shepherd/SKILL.md Rewritten around the shepherd-driver loop + ownership signaling + fold-by-default invariants.
.agents/skills/batch-bug-shepherd/design.md New genesis design record; counterpart to SKILL.md.
.agents/skills/batch-bug-shepherd/assets/shepherd-driver-prompt.md New unified per-PR convergence loop with steps X.0..X.8.
.agents/skills/batch-bug-shepherd/assets/fold-vs-defer-rubric.md New rubric defining scope-creep as the decision axis.
.agents/skills/batch-bug-shepherd/assets/copilot-classification-prompt.md New LEGIT/NOT-LEGIT classification template for Copilot review.
.agents/skills/batch-bug-shepherd/assets/ci-recovery-checklist.md New post-push watch + 4-bucket recovery loop (cap 3).
.agents/skills/batch-bug-shepherd/assets/strategic-alignment-prompt.md New Phase 1.5 ceo-align spawn body; promises schema validation.
.agents/skills/batch-bug-shepherd/assets/conflict-resolution-prompt.md New Phase 5b spawn body; references a comment block name not present in the template.
.agents/skills/batch-bug-shepherd/assets/progress-diagram.md New operator-visibility mermaid; subgraph labels still use pre-refactor phase names.
.agents/skills/batch-bug-shepherd/assets/verdict-schema.json Extends completion_return + new status enum; missing strategic_alignment_return definition the new prompts promise.
.agents/skills/batch-bug-shepherd/assets/final-report-template.md Adds per-PR mergeability row + saga-end aggregated table + folded/deferred sections.
.agents/skills/batch-bug-shepherd/assets/ground-truth-table.md Adds new shepherd-driver-iter-* and advisory-with-deferred statuses.
.agents/skills/batch-bug-shepherd/assets/fix-prompt.md Notes the hand-off to shepherd-driver and CI checklist on first-push red.
.agents/skills/batch-bug-shepherd/references/mergeability-gate.md New Phase 5 reference procedure (5a probe, 5b fan-out, 5c synthesis).
.agents/skills/batch-bug-shepherd/references/strategic-alignment-gate.md New Phase 1.5 reference procedure with fail-open semantics.
.agents/skills/batch-bug-shepherd/assets/shepherd-prompt.md Deleted (absorbed into shepherd-driver).
.agents/skills/batch-bug-shepherd/assets/completion-prompt.md Deleted (absorbed into shepherd-driver).

Copilot's findings

  • Files reviewed: 18/18 changed files
  • Comments generated: 3

Comment thread .agents/skills/batch-bug-shepherd/assets/conflict-resolution-prompt.md
Comment on lines +130 to +131
resolved`. Render from the RESOLUTION CONFIRMATION COMMENT block
in `final-report-template.md`. Include:
Comment thread .agents/skills/batch-bug-shepherd/assets/progress-diagram.md
Comment on lines +53 to +60
subgraph WAVE2[" "]
direction LR
P3a["Phase 3a<br/>shepherd<br/>k = <k> PRs in flight"]:::pending
P3b["Phase 3b<br/>fix dispatch<br/>m = <m> rows without PR"]:::pending
end

P4["Phase 4<br/>completion<br/>F = <F> PRs needing follow-up"]:::pending

Comment thread .agents/skills/batch-bug-shepherd/assets/verdict-schema.json
Comment on lines +119 to +123
"panel_final_verdict": {
"type": "string",
"enum": ["ship_now", "ship_with_followups", "needs_discussion", "needs_rework"],
"description": "CEO stance from the final panel pass in this run."
},
@danielmeppiel
feat(skills/bbs): backfill genesis evals plan + real-task refinement …
c8ddee4 
…evidence

Backfills the genesis Step 6 EVALS PLAN and Step 8 EVALS GATE that
the structural refactor (PR #1518) shipped without. Per genesis the
EVALS GATE blocks any future shipping of skill-body changes.

Adds three structured-input content evals that exercise the load-
bearing decision policies introduced by the v2 refactor:

  * fold-vs-defer-panel  -- Phase X.2 rubric application on a
    panel CEO ship_with_followups return (3 fold, 1 defer with
    scope_boundary_crossed).
  * copilot-classification-and-fold -- Phase X.0 LEGIT/NOT-LEGIT
    classification on 5 inline comments (4 LEGIT folded, 1
    NOT-LEGIT dismissed with rationale).
  * ci-recovery-lint-bucket -- Phase X.6 bucket-1 lint recovery
    via ruff format + push + watch re-entry, cap 3.

Each scenario ships with_skill and without_skill fixtures and a
regex rubric scored by the existing scripts/run_evals.py runner.
All five content scenarios + the trigger val split pass:

  triggers val: 1.0 should-fire, 1.0 should-not-fire
  content delta_anchors: 5, 7, 7, 7, 8 (gate: >=1)

Adds evals/real-task-refinement.md capturing the wave-1 (v1
SKILL.md) vs wave-2 (v2 SKILL.md) comparison that drove the
default-fold + Copilot-first-class + CI-recovery-first-class
edits, plus the iter-4 #1513 CI-infra rollback as a positive
trace of the new bucket-3 recovery path.

Adds an Evals section near the bottom of SKILL.md pointing at
evals/ and stating the EVALS GATE for future skill-body changes.

ASCII guard clean. auth-signals lint clean. Scope: skill-only;
install.ps1 and tests/unit/install/test_windows_shim_template.py
remain untracked (they belong to PR #1512, not here).

Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
@danielmeppiel danielmeppiel merged commit fbf3b06 into main May 27, 2026
19 checks passed
@danielmeppiel danielmeppiel deleted the feat/bbs-shepherd-recommendation-fold-loop branch May 27, 2026 22:15
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Reviewers

Copilot code review Copilot Copilot left review comments

@sergio-sisternes-epam sergio-sisternes-epam Awaiting requested review from sergio-sisternes-epam sergio-sisternes-epam is a code owner

Assignees

@danielmeppiel danielmeppiel

Labels

None yet

Projects

None yet

Milestone

No milestone

Development

Successfully merging this pull request may close these issues.

2 participants

@danielmeppiel