Use when completing tasks, requesting review, or receiving code review feedback — covers both requesting and responding to reviews with technical rigor
From superpower-routernpx claudepluginhub charliechan53/superpower-router --plugin superpower-routerThis skill uses the workspace's default tool permissions.
code-reviewer.mdGuides Next.js Cache Components and Partial Prerendering (PPR) with cacheComponents enabled. Implements 'use cache', cacheLife(), cacheTag(), revalidateTag(), static/dynamic optimization, and cache debugging.
Migrates code, prompts, and API calls from Claude Sonnet 4.0/4.5 or Opus 4.1 to Opus 4.5, updating model strings on Anthropic, AWS, GCP, Azure platforms.
Details PluginEval's skill quality evaluation: 3 layers (static, LLM judge), 10 dimensions, rubrics, formulas, anti-patterns, badges. Use to interpret scores, improve triggering, calibrate thresholds.
Covers both sides of code review: requesting reviews (dispatching reviewer subagents) and receiving feedback (verifying before implementing suggestions).
After completing implementation, dispatch two reviews in order:
Route to Codex CLI (read-only):
${CLAUDE_PLUGIN_ROOT}/skills/plan-and-execute/codex-runner.sh \
"Review code for spec compliance. Spec: [requirements]. Files: [paths]. Report: APPROVED or NEEDS CHANGES with specifics." \
read-only /path/to/project
Fail-closed: if Codex returns exit 20, ask user for explicit approval before any Claude/Sonnet fallback.
Optional fallback only when CODEX_FAIL_CLOSED=0: Task(prompt, model:"sonnet", subagent_type:"superpower-router:code-reviewer")
Only after spec compliance passes. Route to Codex CLI (read-only):
${CLAUDE_PLUGIN_ROOT}/skills/plan-and-execute/codex-runner.sh \
"Review code quality. Files: [paths]. Check: naming, error handling, DRY, YAGNI, test coverage. Report: Strengths, Issues (Critical/Important/Minor), APPROVED or NEEDS CHANGES." \
read-only /path/to/project
If reviewer finds issues → fix → re-review → repeat until approved.
When YOU receive feedback, apply technical rigor:
| Step | Action |
|---|---|
| 1 | Read feedback carefully — don't skim |
| 2 | Verify each suggestion is technically correct before implementing |
| 3 | If suggestion seems wrong, research it (WebSearch, docs) before disagreeing |
| 4 | If genuinely wrong, explain why with evidence — don't just agree |
| 5 | Implement valid suggestions, explain rejections |
See ./code-reviewer.md for the detailed reviewer agent prompt template.