paritytech/trinity-user-agents/reviewer
Claude Code and agent skill by paritytech/trinity-user-agents
Trust tier 3
- community
Source check pending
Checking that the SKILL.md file is still in the repository.
Review a PR or diff against architecture and code checklists. Produces a verdict with blocking/major/minor findings.
Install
npx vskill@latest install paritytech/trinity-user-agents/reviewerSKILL.md
Reviewer Skill
Procedure
Step 1: Get the Diff
Determine what to review:
- If a PR number is provided:
gh pr diff {number} --repo paritytech/polkadot-app-ios-v2 - If reviewing local changes:
git difforgit diff develop...HEAD - Read the PR description if available:
gh pr view {number} --repo paritytech/polkadot-app-ios-v2
Step 2: Read PLAN.md (if present)
If .claude/PLAN.md exists and relates to this PR:
- Read it fully
- Note
files_touched,must_not_touch,out_of_scope - Will check plan-vs-implementation alignment in Step 5
Step 3: Load Checklists
Load the review checklists as the primary review source:
.claude/docs/review/architecture-checklist.md— for structural/design review.claude/docs/review/code-checklist.md— for file-by-file code review
Step 4: Walk the Diff
For each changed file:
- Read the full file (not just the diff) to understand context
- Apply relevant checklist items
- Note any violations with severity:
- Blocking — must fix before merge (correctness, security, data loss risk)
- Major — should fix before merge (architecture violation, maintainability)
- Minor — nice to fix (style, naming, minor improvement)
Group findings by theme, not by file.
Step 5: Plan-vs-Implementation Diff (if PLAN.md exists)
Check:
- [ ] All
files_touchedin the plan were actually modified - [ ] No files in
must_not_touchwere modified - [ ] Nothing from
out_of_scopewas implemented - [ ] The approach matches the plan's design decisions
- [ ] Verification criteria from the plan are addressed
Step 6: Doc-Update Proposals
Check if the changes reveal:
- Code-level gaps — patterns used that aren't in
code/docs - Architecture-level gaps — new architectural decisions not in
architecture/docs - Checklist gaps — rules that should be added to review checklists
Propose specific doc updates if any gaps found.
Step 7: Write Verdict
Format the review output:
Inline Comment Style
When posting GitHub comments:
- Tag severity:
[blocking],[major],[minor] - Quote
file:line - Cite the checklist rule
- Suggest a concrete fix
- 1-2 paragraphs max, no fluff
Summary Body (for PR comment)
- Lead with overall sentiment (approve / request changes)
- List cross-cutting themes (not per-file noise)
- Process notes only if relevant (e.g., "plan says X but implementation does Y")
- Keep it actionable
Important Notes
- The reviewer does NOT extract rules — only architect and implementer do
- If no PLAN.md exists, review against checklists only
- When in doubt about a rule, cite the specific doc section
- Focus on what matters: correctness > architecture > style
Security
- Security scanPassed, 95 of 100Pattern checks on the skill's files.
- Model reviewPassed
Quality evaluation
- Task evaluationNot runNo evaluation yet. Run one in Skill Studio.
Source
- SKILL.md in the repositoryCheckinghosts/ios/.claude/plugins/polkadotapp-workflows/skills/reviewer/SKILL.md
Popularity
- 7-day trend score
- 0
Not refreshed yet.
Works with
Any agent vskill installs to: 56 targets in all. These 39 have a page on this site.
- Claude Code
- Cursor
- GitHub Copilot
- Windsurf
- Codex
- Gemini CLI
- Cline
- Amp
- Roo Code
- Goose
- Continue
- Kilo Code
- Junie
- Trae
- OpenHands
- Qwen Code
- AI Agent
- Tabnine Chat
- Sourcegraph Cody
- Amazon Q
- Firebase Genkit
- Sweep
- Devin
- Phind
- Mentat
- Claude Desktop
- Aider
- GPT Pilot
- Plandex
- bolt.new
- v0
- Replit Agent
- Anse
- TypingMind
- BetterChatGPT
- LibreChat
- Warp AI
- Pieces
- Fig AI
Embed badge
[](https://verified-skill.com/skills/paritytech/trinity-user-agents/reviewer)