OR

orchestrate:review

Perform standardized checks on orchestration PRs and coordinate their final approval.

Install

mkdir -p .claude/skills/orchestrate-review-kagenti && curl -L -o skill.zip "https://agentskills.codes/api/skills/download/14687" && unzip -o skill.zip -d .claude/skills/orchestrate-review-kagenti && rm skill.zip

Installs to .claude/skills/orchestrate-review-kagenti

Activation

This is the description your AI agent reads to decide when to run this skill — the better it matches your request, the more reliably it fires.

Review all orchestration PRs before merge - per-PR checks, cross-PR consistency, and coordinated approval
105 charsno explicit “when” trigger
Advanced

Key capabilities

  • Gather open orchestration PRs and their metadata
  • Perform per-PR review based on commit conventions, PR format, and area-specific checks
  • Conduct cross-PR consistency checks (e.g., pre-commit ↔ CI alignment)
  • Draft a review summary with verdicts and issues found
  • Submit reviews via GitHub API after user approval

How it works

The skill gathers open orchestration PRs and their metadata. It then performs per-PR reviews against checklists and conducts cross-PR consistency checks. A review summary is drafted for user approval, and reviews are submitted via the GitHub API.

Inputs & outputs

You give it
A request to review all orchestration PRs, e.g., "Review open orchestration PRs."
You get back
A drafted review summary for orchestration PRs, including per-PR verdicts, issues found, and cross-PR consistency checks, with reviews submitted to GitHub upon

When to use orchestrate:review

  • Review open orchestration PRs
  • Check consistency across orchestration changes
  • Approve orchestrated work

About this skill

flowchart TD
    START(["/orchestrate:review"]) --> GATHER["List open orchestration PRs"]:::orch
    GATHER --> PER_PR["Per-PR review"]:::orch
    PER_PR --> CROSS["Cross-PR consistency checks"]:::orch
    CROSS --> DRAFT["Draft review summary"]:::orch
    DRAFT --> APPROVE{User approves?}
    APPROVE -->|Yes| SUBMIT["Post reviews via gh api"]:::orch
    APPROVE -->|No| REVISE["Revise reviews"]:::orch
    REVISE --> DRAFT
    SUBMIT --> STATUS["Update phase-status.md"]:::orch
    STATUS --> DONE([Review complete])

    classDef orch fill:#FF9800,stroke:#333,color:white
    classDef check fill:#FFC107,stroke:#333,color:black
    class APPROVE check

Follow this diagram as the workflow.

Orchestrate: Review

Phase 7 quality gate. Review all orchestration PRs created by phases 2-6 before merge. Checks each PR individually, then validates cross-PR consistency, and submits reviews after user approval.

When to Use

  • After all orchestration phases (2-6) have created their PRs
  • Before merging any orchestration PRs into the target repo
  • When /orchestrate review is invoked from the router

Prerequisites

  • scan-report.md and plan.md exist in /tmp/rossoctl/orchestrate/<target>/
  • Phases 2-6 are complete (or at least the phases that were planned)
  • PRs are open on the target repo

Phase 1: Gather

List open PRs for the target repo and collect metadata:

# List open PRs created by orchestration (look for orchestrate-related branch names or labels)
gh pr list --repo <org>/<repo> --state open --json number,title,headRefName,additions,deletions,files

For each PR, fetch the diff:

gh pr diff <number> --repo <org>/<repo>

Record PR metadata in a working table:

PRTitleBranchFiles+/-
#N...orchestrate/...N+X/-Y

Phase 2: Per-PR Review

For each PR, run the github:pr-review checklist:

Commit Conventions

  • Signed-off (Signed-off-by: trailer present)
  • Emoji prefix on commit message (if repo convention requires it)
  • Imperative mood in subject line
  • Body explains "why" not just "what"

PR Format

  • Title under 70 characters
  • Summary section in PR body
  • Links to relevant issues or plan

Area-Specific Checks

PR PhaseChecks
precommit (Phase 2).pre-commit-config.yaml valid YAML, hooks match detected languages, no conflicting formatters
tests (Phase 3)Test files follow naming conventions, fixtures are reusable, no hardcoded secrets in tests
ci (Phase 4)Actions SHA-pinned, permissions least-privilege, no secrets in logs, workflows valid YAML
security (Phase 5)CODEOWNERS paths exist, SECURITY.md has contact info, LICENSE matches repo intent
replicate (Phase 6)Skills have frontmatter, SKILL.md files are valid markdown, paths reference target repo correctly

Security Review

  • No secrets, tokens, or credentials in diff
  • No overly permissive file permissions
  • No eval, exec, or injection-prone patterns in scripts
  • Container images use specific tags (not :latest)

Phase 3: Cross-PR Consistency

Check alignment across all orchestration PRs:

Pre-commit ↔ CI Alignment

  • Linters configured in .pre-commit-config.yaml (Phase 2) should match lint steps in CI workflows (Phase 4)
  • Example: if pre-commit runs ruff, CI should also run ruff (or at least not run a conflicting linter like flake8)
# Extract pre-commit hooks
grep "repo:\|id:" .repos/<target>/.pre-commit-config.yaml 2>/dev/null
# Compare with CI lint steps
grep -A5 "lint\|check\|format" .repos/<target>/.github/workflows/*.yml 2>/dev/null

Tests ↔ CI Alignment

  • Tests added in Phase 3 should be executed by CI workflows added in Phase 4
  • Check that test commands in CI match the test framework detected
# Test framework from Phase 3
grep -r "pytest\|go test\|vitest\|jest" .repos/<target>/.github/workflows/*.yml 2>/dev/null

CODEOWNERS ↔ Paths

  • Paths in CODEOWNERS (Phase 5) should cover directories created by earlier phases
# Check CODEOWNERS paths exist
cat .repos/<target>/CODEOWNERS 2>/dev/null | grep -v "^#" | awk '{print $1}' | while read path; do
  ls .repos/<target>/$path 2>/dev/null || echo "MISSING: $path"
done

Skills ↔ Repo Paths

  • Skills replicated in Phase 6 should reference correct paths for the target repo
  • Skill frontmatter should be valid
# Check skill files have valid frontmatter
find .repos/<target>/.claude/skills -name "SKILL.md" -exec head -5 {} \; 2>/dev/null

Phase 4: Draft

Present a review summary to the user. Format:

# Orchestration Review: <target>

## Per-PR Verdicts

| PR | Title | Verdict | Issues |
|----|-------|---------|--------|
| #N | precommit: ... | approve | 0 |
| #N | tests: ... | request-changes | 2 |
| #N | ci: ... | approve | 0 |
| #N | security: ... | comment | 1 |
| #N | replicate: ... | approve | 0 |

## Issues Found

### PR #N: <title>
1. **[severity]** Description of issue
   - File: `path/to/file`
   - Recommendation: ...

## Cross-PR Consistency

| Check | Status | Notes |
|-------|--------|-------|
| Pre-commit ↔ CI lint | aligned/misaligned | details |
| Tests ↔ CI execution | aligned/misaligned | details |
| CODEOWNERS ↔ paths | aligned/misaligned | details |
| Skills ↔ repo paths | aligned/misaligned | details |

Present this to the user and wait for approval before submitting.

Phase 5: Submit

After user approval, post reviews via GitHub API:

# For each PR, post the review
gh api repos/<org>/<repo>/pulls/<number>/reviews \
  --method POST \
  -f event="APPROVE" \
  -f body="Orchestration review: all checks passed. ..."

# Or for request-changes:
gh api repos/<org>/<repo>/pulls/<number>/reviews \
  --method POST \
  -f event="REQUEST_CHANGES" \
  -f body="Orchestration review: issues found. ..."

For PRs with inline comments, use the review comments API:

gh api repos/<org>/<repo>/pulls/<number>/reviews \
  --method POST \
  -f event="REQUEST_CHANGES" \
  -f body="..." \
  --input comments.json

Where comments.json contains file-level comments.

Status Update

Update phase-status.md when complete:

# Update phase-status.md
sed -i '' 's/| review .*/| review | complete | -- | YYYY-MM-DD |/' /tmp/rossoctl/orchestrate/<target>/phase-status.md

Related Skills

  • orchestrate -- Parent router
  • github:pr-review -- Per-PR review checklist (invoked during Phase 2)
  • orchestrate:scan -- Scan report used for cross-referencing
  • orchestrate:plan -- Plan used to verify all phases were executed

When not to use it

  • Before all orchestration phases (2-6) have created their PRs
  • When the user wants to merge orchestration PRs without review
  • When the user is asking for general code review not specific to orchestration PRs

Prerequisites

`scan-report.md` exists in `/tmp/kagenti/orchestrate/<target>/``plan.md` exists in `/tmp/kagenti/orchestrate/<target>/`Phases 2-6 are complete (or at least the phases that were planned)PRs are open on the target repo

Limitations

  • Requires `scan-report.md` and `plan.md` to exist.
  • Assumes PRs are open on the target repo.
  • The skill relies on `gh` commands for GitHub API interaction.

How it compares

This skill automates the review process for orchestration PRs by performing individual and cross-PR consistency checks, drafting a summary, and submitting reviews via the GitHub API, providing a structured and verifiable quality gate compar

Compared to similar skills

orchestrate:review side by side with the closest alternatives in the catalog.

SkillInstallsUpdatedSafetyDifficulty
orchestrate:review (this skill)05moReviewAdvanced
resolve-conflicts818moReviewIntermediate
claude-automation-recommender472moReviewBeginner
codex-skill125moReviewAdvanced

Try saying

Example prompts that trigger this skill in your AI assistant.

You might also like

resolve-conflicts

antinomyhq

Use this skill immediately when the user mentions merge conflicts that need to be resolved. Do not attempt to resolve conflicts directly - invoke this skill first. This skill specializes in providing a structured framework for merging imports, tests, lock files (regeneration), configuration files, and handling deleted-but-modified files with backup and analysis.

81334

claude-automation-recommender

anthropics

Analyze a codebase and recommend Claude Code automations (hooks, subagents, skills, plugins, MCP servers). Use when user asks for automation recommendations, wants to optimize their Claude Code setup, mentions improving Claude Code workflows, asks how to first set up Claude Code for a project, or wants to know what Claude Code features they should use.

47140

codex-skill

feiskyer

Use when user asks to leverage codex, gpt-5, or gpt-5.1 to implement something (usually implement a plan or feature designed by Claude). Provides non-interactive automation mode for hands-off task execution without approval prompts.

12110

git-advanced-workflows

wshobson

Master advanced Git workflows including rebasing, cherry-picking, bisect, worktrees, and reflog to maintain clean history and recover from any situation. Use when managing complex Git histories, collaborating on feature branches, or troubleshooting repository issues.

1197

subagent-driven-development

davila7

Use when executing implementation plans with independent tasks in the current session

1493

validate-openapi-specs

epieczko

Validates and registers hook manifest files (YAML) in the Hook Registry for versioned hook management.

594

Search skills

Search the agent skills registry