Menu
Use when implementation is complete and a comprehensive code review is needed — dispatches parallel review agents and merges findings
Install with Codex or Claude Copy this prompt, paste it into Codex, Claude, or another assistant, and let it review the skill page and install it for you.
Based on SOC occupation classification
Use when a structured PRD document is needed for the ralph loop — generates prd.json and progress.txt
Use when task scope exceeds one context window and decomposes into independent pass/fail stories with clean iteration boundaries
Use when implementation is complete, all tests pass, and you need to decide how to integrate the work — merge, PR, or cleanup
Use when starting feature work that needs isolation from the current workspace or before executing plans that could leave the branch in a broken state
Use when a hard problem was just solved and the solution should be captured as a learning for future reference
Use when 3+ distinct occurrences suggest a skill should be updated with a new trigger, anti-pattern, or verification step
| name | requesting-code-review |
| description | Use when implementation is complete and a comprehensive code review is needed — dispatches parallel review agents and merges findings |
| tags | ["review","quality","agents"] |
| triggers | ["code review","review this","review my code","PR review","request review"] |
| chains_to | ["verification-before-completion"] |
| priority | core |
| gate | false |
Dispatch parallel review agents against the current diff. Each agent reviews from a different angle. Merge findings, deduplicate, prioritize by severity, and present a unified review.
self-review when the sensei suggests deeper reviewGet the diff that will be reviewed:
# For unstaged changes
git diff
# For staged changes
git diff --staged
# For branch comparison
git diff main...HEAD
Based on the changes, pick the relevant agents:
| Change Type | Agents to Dispatch |
|---|---|
| Any code change | simplicity-reviewer (always) |
| New features, refactors | architecture-reviewer + simplicity-reviewer |
| API endpoints, auth, user input | security-reviewer + architecture-reviewer |
| Database queries, hot paths, render loops | performance-reviewer |
| Files matching registered patterns | pattern-reviewer |
| Large changes (100+ lines, 5+ files) | All review agents |
Minimum: simplicity-reviewer (always). Maximum: all 5 (for large changes).
Launch selected agents in a single message using the Agent tool:
Agent 1 (simplicity-reviewer): Review this diff for over-engineering and unnecessary code
Agent 2 (security-reviewer): Review this diff for security vulnerabilities
Agent 3 (pattern-reviewer): Check this diff against registered patterns
Each agent prompt should include:
After all agents return:
CODE REVIEW: <one-line summary>
P1 (must fix before merge):
- file:line — [reviewer] — [finding + fix]
P2 (should fix):
- file:line — [reviewer] — [finding + fix]
P3 (nice to have):
- file:line — [reviewer] — [finding]
Reviewers: simplicity, security, pattern (3 dispatched)
After fixes, re-run verification → chain to verification-before-completion.
| Anti-Pattern | Why It's Wrong |
|---|---|
| Reviewing before tests pass | Fix functional issues first, then review quality |
| Dispatching all 5 agents for a 10-line change | Overkill. Sensei self-review is sufficient. |
| Ignoring P1 findings | P1 means "must fix." No exceptions. |
| Running agents sequentially | Dispatch in parallel. It's faster. |
| Not deduplicating findings | Presenting the same issue 3 times wastes the user's time. |
REQUIRED: syntaxninja-dojo:verification-before-completion (after fixes are applied)