技能评估器

skill-reviewer

by ann0501

Score and grade AgentSkills across four dimensions: design quality, content quality, security, and usability. Use when asked to score, grade, evaluate, or rate a skill's quality — producing a numeric scorecard with dimension-by-dimension ratings. Triggers on "score this skill", "grade skill", "rate skill quality", "evaluate skill", "skill scorecard", "给技能打分", "技能评分".

4.5k内容与创意未扫描2026年4月20日

安装

claude skill add --url https://github.com/openclaw/skills

文档

Skill Reviewer

Four-dimension quality auditor for AgentSkills.

Review Process

Step 1: Inventory & Classify

  1. find /path/to/skill -type f — list all files. If path doesn't exist or is empty, abort with error.
  2. Read _meta.json or .clawhub/origin.json if present (source, version, author info).
  3. Read SKILL.md in full.
  4. Classify skill type per references/weight-adjustment.md → determine weights.

Step 2: Scan

  1. SKILL.md (full read)
  2. scripts/ (full read, do NOT execute)
  3. references/ (full read if ≤10 files or ≤50KB total; spot-check otherwise)
  4. assets/ (list filenames, flag non-asset file types: .sh/.py/.exe/.bat/.ps1 in assets/ = suspicious)
  5. Other files (.json, dotfiles, symlinks) — inspect each

Step 3: Score Each Dimension

Load reference files: references/design-quality.md, references/content-quality.md, references/security.md, references/usability.md.

For each dimension, walk the checklist and score per the rubric. Cite [file:line] for each finding.

Security shortcut: Run grep sweeps for CRITICAL patterns first (see security.md C1-C11). If any CRITICAL found → D3=1, note it, but still score remaining dimensions — don't skip them.

Orphan file check (security.md): After reading SKILL.md, cross-reference all mentioned files against the actual file list. Flag orphans.

Step 4: Calculate & Verdict

Weighted total = Σ(dim_score × adjusted_weight)

VerdictCondition
✅ PASSTotal ≥ 8 AND every dimension ≥ 6 AND no CRITICAL findings
🔧 REVISETotal < 8 OR any dimension < 6 OR any CRITICAL finding

CRITICAL auto-revise: Even if total ≥ 8, any security CRITICAL finding forces REVISE.

Report Format

Use list format (not tables — compatible with all platforms):

PASS:

code
【技能审校通过】
技能:{name}
路径:{path}
类型:{type}
综合得分:{score}/10
审校日期:{date}

• 设计质量({w1}%):{d1}/10 — {简要说明}
• 内容质量({w2}%):{d2}/10 — {简要说明}
• 安全性({w3}%):{d3}/10 — {简要说明}
• 实用性({w4}%):{d4}/10 — {简要说明}

亮点:
- {亮点1}
- {亮点2}

REVISE:

code
【技能审校打回】
技能:{name}
路径:{path}
类型:{type}
综合得分:{score}/10
审校日期:{date}

⚠️ 判定维度:
• 设计质量({w1}%):{d1}/10
• 内容质量({w2}%):{d2}/10
• 安全性({w3}%):{d3}/10
• 实用性({w4}%):{d4}/10

具体问题:
- [D1] {issue} — {file}:{line}
- [D3] 🔴 CRITICAL: {issue} — {file}:{line}

改进方向:
- {方向1}
- {方向2}

达标标准:
- {标准1}
- {标准2}

References

  • references/weight-adjustment.md — type classification + weight matrix
  • references/design-quality.md — D1 checklist + rubric + scoring rules
  • references/content-quality.md — D2 checklist + rubric + scoring rules
  • references/security.md — D3 pattern table + detection commands
  • references/usability.md — D4 checklist + rubric + scoring rules

相关 Skills

文档共著

by anthropics

Universal
热门

围绕文档、提案、技术规格、决策记录等写作任务,按上下文收集、结构迭代、读者测试三步协作共创,减少信息遗漏,写出更清晰、经得起他人阅读的内容。

写文档、方案或技术规格时容易思路散、信息漏,它用结构化共著流程帮你高效传递上下文、反复打磨内容,还能从读者视角做验证。

内容与创意
未扫描139.0k

内部沟通

by anthropics

Universal
热门

按公司常用模板和语气快速起草内部沟通内容,覆盖 3P 更新、状态报告、领导汇报、项目进展、事故复盘、FAQ 与 newsletter,适合需要统一格式的团队沟通场景。

按公司偏好的模板快速产出状态汇报、领导更新和 FAQ,既省去反复改稿,也让内部沟通更统一、更专业。

内容与创意
未扫描139.0k

平面设计

by anthropics

Universal
热门

先生成视觉哲学,再落地成原创海报、艺术画面或其他静态设计,输出 .png/.pdf,强调构图、色彩与空间表达,适合需要高完成度视觉成品的场景。

做海报、插画或静态视觉稿时,用它能快速产出兼顾美感与版式的PNG/PDF成品,原创设计更省心,也更适合规避版权风险。

内容与创意
未扫描139.0k

相关 MCP 服务

免费的加密新闻聚合 MCP,汇集 Bitcoin、Ethereum、DeFi、Solana 与 altcoins 资讯源。

内容与创意
212

by ProfessionalWiki

热门

让 Large Language Model 客户端无缝连接任意 MediaWiki 站点,可创建、更新、搜索页面,并通过 OAuth 2.0 安全管理内容。

内容与创意16 个工具
94

by roomi-fields

热门

Automate Google NotebookLM — Q&A with citations, audio, video, content generation

内容与创意
79

评论