用 Codex 或 Claude 帮你安装 复制这段 Prompt,粘贴到 Codex、Claude 或其他助手里,让它检查 Skill 页面并帮你完成安装。
直接命令不会经过审查 Prompt;运行前请先检查来源。
npx skills add https://github.com/diegosouzapw/awesome-omni-skill --skill preen-review-instructions命令会保持在同一行。复制前请横向滚动并检查完整内容。
想先保存到本地?可下载 SkillsMP 当前能够提供的文件。
Token-efficient tracking for AI orchestration. CLI-first for status updates (~50 tokens), agent fallback for complex ops (~1KB). Use when: updating task status, querying blockers, creating progress files, validating phases.
AshAi extension guidelines for integrating AI capabilities with Ash Framework. Use when implementing vectorization/embeddings, exposing Ash actions as LLM tools, creating prompt-backed actions, or setting up MCP servers. Covers semantic search, LangChain integration, and structured outputs.
This skill should be used when solving hard questions, complex architectural problems, or debugging issues that benefit from GPT-5 Pro or GPT-5.1 thinking models with large file context. Use when standard Claude analysis needs deeper reasoning or extended context windows.
正在显示 SKILL.md
基于 SOC 职业分类
| name | preen-review-instructions |
| description | Audit and update code review instructions (REVIEW.md, .gemini/INSTRUCTIONS.md) |
Audit the review instructions in REVIEW.md and .gemini/INSTRUCTIONS.md for completeness, accuracy, and sync with codebase patterns. Update instructions to reflect current conventions.
Run this skill when:
| File | Purpose |
|---|---|
REVIEW.md | Master review instructions (repo root) |
.gemini/INSTRUCTIONS.md | Gemini Code Assist specific instructions |
CLAUDE.md | Claude Code conventions (source of truth) |
AGENTS.md | Codex conventions (source of truth) |
compliance/ | Compliance documentation (security controls) |
compliance/infrastructure-controls.md | Cross-framework sentinel index |
echo "=== REVIEW.md sections ==="
rg '^##' REVIEW.md | head -20
echo "=== Last modified ==="
git log -1 --format="%ci" -- REVIEW.md
# New route files not mentioned
echo "=== Recent route additions ==="
git log --since="30 days ago" --name-only --pretty=format: -- 'packages/api/src/routes/**/*.ts' | sort -u | head -10
# New shared components not mentioned
echo "=== Recent component additions ==="
git log --since="30 days ago" --name-only --pretty=format: -- 'packages/*/src/components/**/*.tsx' | sort -u | head -10
# New package directories
echo "=== Package structure ==="
ls -d packages/*/ 2>/dev/null | head -10
# TypeScript rules
echo "=== CLAUDE.md TypeScript mentions ==="
rg -i 'typescript|any|cast|@ts-' CLAUDE.md | head -10
# Security rules
echo "=== CLAUDE.md security mentions ==="
rg -i 'security|auth|permission|injection' CLAUDE.md | head -10
echo "=== Gemini INSTRUCTIONS sections ==="
rg '^##' .gemini/INSTRUCTIONS.md | head -20
# Compare section counts
echo "=== Section comparison ==="
echo "REVIEW.md sections: $(rg '^##' REVIEW.md | wc -l)"
echo "Gemini sections: $(rg '^##' .gemini/INSTRUCTIONS.md | wc -l)"
echo "=== Review script prompt analysis ==="
rg -A5 'PROMPT=' scripts/solicitClaudeCodeReview.sh | head -15
rg -A5 'PROMPT=' scripts/solicitCodexReview.sh 2>/dev/null | head -15 || true
# Check REVIEW.md has security section
echo "=== REVIEW.md security coverage ==="
rg -c 'Security|OWASP|injection|auth' REVIEW.md
# Check for compliance documentation references
echo "=== Compliance references in REVIEW.md ==="
rg -c 'compliance/|sentinel|TL-' REVIEW.md
# Check infrastructure-controls.md exists and has sentinels
echo "=== Infrastructure controls sentinel count ==="
rg -c 'TL-[A-Z]+-[0-9]+' compliance/infrastructure-controls.md
# Check for framework parity
echo "=== Framework document counts ==="
for fw in HIPAA NIST.SP.800-53 SOC2; do
echo "$fw policies: $(ls compliance/$fw/policies/*.md 2>/dev/null | wc -l | tr -d ' ')"
done
# Check REVIEW.md mentions sentinel workflow
echo "=== Sentinel workflow in REVIEW.md ==="
rg 'Adding New Sentinel|COMPLIANCE_SENTINEL' REVIEW.md | head -5
| Category | Severity | Action |
|---|---|---|
| Missing section for new pattern | Medium | Add section to REVIEW.md, sync to Gemini |
| Outdated package structure | Low | Update package list in REVIEW.md |
| Gemini instructions drift | Medium | Sync .gemini/INSTRUCTIONS.md with REVIEW.md |
| Security rule missing | High | Add security guidance from CLAUDE.md |
| Review script not using instruct. | Medium | Update script to include REVIEW.md content |
| Missing compliance section | High | Add compliance documentation guidance |
| Missing sentinel workflow | High | Add instructions for adding new sentinels |
| Compliance framework parity gap | Medium | Ensure all frameworks have equal coverage |
REVIEW.md under appropriate heading.gemini/INSTRUCTIONS.mdREVIEW.md# Compare section headers
diff <(rg '^##' REVIEW.md | sort) <(rg '^##' .gemini/INSTRUCTIONS.md | sort)
Ensure key sections from REVIEW.md have corresponding entries in .gemini/INSTRUCTIONS.md.
If solicitClaudeCodeReview.sh doesn't reference REVIEW.md:
# Read and include instructions
INSTRUCTIONS=$(cat REVIEW.md | head -100)
PROMPT="Review using these guidelines:
$INSTRUCTIONS
[Rest of prompt...]"
When security or compliance guidance is missing from review instructions:
/preen-api-security skill for security patterns to documentcompliance/infrastructure-controls.md for sentinel patternsRequired content for security section:
Required content for compliance section:
TL-<CATEGORY>-<NUMBER>)infrastructure-controls.mdWhen new security controls need compliance documentation:
Add sentinel to compliance/infrastructure-controls.md with:
Create/update document triad in each framework:
compliance/<FW>/policies/NN-topic-policy.mdcompliance/<FW>/procedures/NN-topic-procedure.mdcompliance/<FW>/technical-controls/NN-topic-control-map.mdUpdate POLICY_INDEX.md in each framework directory
Add COMPLIANCE_SENTINEL comments in policy documents:
<!-- COMPLIANCE_SENTINEL: TL-NEW-001 | policy=path | procedure=path | control=description -->
git checkout -b docs/review-instructions-<date>/commit-and-push, then /enter-merge-queueIf no issues found during discovery, do not create a branch.
.gemini/INSTRUCTIONS.md concise (Gemini has token limits)REVIEW.md and .gemini/INSTRUCTIONS.mdREVIEW.md.gemini/INSTRUCTIONS.md covers same topics as REVIEW.md (condensed)compliance/infrastructure-controls.md and /preen-compliance-docsThe quality metric is the count of missing or stale sections:
# Count potential gaps
GAPS=0
# Check for missing sections
[ -z "$(rg 'TypeScript Standards' REVIEW.md)" ] && GAPS=$((GAPS + 1))
[ -z "$(rg 'API Security' REVIEW.md)" ] && GAPS=$((GAPS + 1))
[ -z "$(rg 'React Standards' REVIEW.md)" ] && GAPS=$((GAPS + 1))
[ -z "$(rg 'Database Performance' REVIEW.md)" ] && GAPS=$((GAPS + 1))
[ -z "$(rg 'Testing Standards' REVIEW.md)" ] && GAPS=$((GAPS + 1))
# Check for security and compliance coverage
[ -z "$(rg 'Security and Compliance' REVIEW.md)" ] && GAPS=$((GAPS + 1))
[ -z "$(rg 'OWASP' REVIEW.md)" ] && GAPS=$((GAPS + 1))
[ -z "$(rg 'Adding New Sentinel' REVIEW.md)" ] && GAPS=$((GAPS + 1))
[ -z "$(rg 'COMPLIANCE_SENTINEL' REVIEW.md)" ] && GAPS=$((GAPS + 1))
[ -z "$(rg 'infrastructure-controls.md' REVIEW.md)" ] && GAPS=$((GAPS + 1))
# Check gemini security coverage
[ -z "$(rg 'Security' .gemini/INSTRUCTIONS.md)" ] && GAPS=$((GAPS + 1))
[ -z "$(rg 'compliance/' .gemini/INSTRUCTIONS.md)" ] && GAPS=$((GAPS + 1))
# Check gemini sync
REVIEW_SECTIONS=$(rg '^## ' REVIEW.md | wc -l)
GEMINI_SECTIONS=$(rg '^## ' .gemini/INSTRUCTIONS.md | wc -l)
[ -lt $((REVIEW_SECTIONS / )) ] && GAPS=$((GAPS + ))
# Limit discovery output
git log --since="30 days ago" --name-only --pretty=format: | head -30
# Suppress validation output
git commit -S -m "message" >/dev/null
git push >/dev/null
On failure, re-run without suppression to see errors.