Skills skill-optimizer
install
source · Clone the upstream repo
git clone https://github.com/wulaosiji/skills
Claude Code · Install into ~/.claude/skills/
T=$(mktemp -d) && git clone --depth=1 https://github.com/wulaosiji/skills "$T" && mkdir -p ~/.claude/skills && cp -r "$T/skill-optimizer" ~/.claude/skills/wulaosiji-skills-skill-optimizer && rm -rf "$T"
manifest:
skill-optimizer/SKILL.mdsource content
Skill Optimizer
Analyze, audit, and optimize skills for maximum discoverability and agent routing accuracy.
When to Use
Use this skill when:
- A user wants to improve an existing skill's discoverability or search ranking
- A skill is not being found by the agent when it should be triggered
- A skill's description is unclear, lacks keywords, or has poor boundaries
- A skill has ambiguous naming that conflicts with other skills
- A user wants to SEO-optimize their skill collection
- Creating a new skill and want to follow best practices from the start
- A skill needs better guardrails to prevent mis-invocation
Do NOT use this skill if:
- The skill is working perfectly and has high usage — don't fix what isn't broken
- The user wants to add new functionality — use feature development instead
- The issue is a security vulnerability or bug in the skill's code — use skill-security-audit instead
Typical triggers:
- 「帮我优化这个skill」「SEO优化技能」「技能描述怎么改」
- "skill search不到", "agent找不到我的技能", "skill命名有问题"
- "重构skill描述", "技能发现率低", "提升skill使用率"
- "optimize my skill", "SEO for skills", "improve skill discoverability"
Workflow
Step 1: Load and Analyze Target Skill
Ask the user for:
- The skill name or path to SKILL.md they want to optimize
- Any specific issues they've noticed (optional)
Load the
SKILL.md file and analyze its current state.
Step 2: Comprehensive Audit
Analyze the skill across 7 dimensions:
- Naming Audit — kebab-case? No underscores/camelCase? Descriptive and discoverable?
- Description Audit — Trigger keywords in multiple languages? "Use when..." explicit? Cross-references included? Output format stated?
- When to Use Boundaries — Clear conditions? Anti-patterns listed? Typical trigger phrases documented?
- Cross-Reference Network — Bidirectional references to related skills? Suite/kit skills linked?
- Guardrails & Constraints — Explicit anti-patterns? Output constraints? Dependencies handled?
- README Integration — Listed in hub README? Correct category? Compelling one-liner?
- Technical SEO — Relevant YAML frontmatter? Correct file structure? Example usage patterns?
Step 3: Generate Optimization Report
Create a structured report with these sections:
<optimized-description> ```## Skill Audit Report: <skill-name> ### Overall Score: X/100 ### 1. Naming (Score: X/10) **Current**: `<current-name>` **Status**: ✅ Good | ⚠️ Needs Improvement | ❌ Critical Issue **Recommendation**: `<suggested-name>` ### 2. Description (Score: X/20) **Current**: `<current-description>` **Optimized Version**:
3. When to Use Boundaries (Score: X/20)
Missing Elements: <list> Recommended Addition: <snippet>
4. Cross-References (Score: X/15)
Recommended Cross-References:
— <reason><related-skill>
5. Guardrails (Score: X/15)
Recommended Additions: <list>
6. README Integration (Score: X/10)
Status: ✅ | ⚠️ | ❌
7. Technical SEO (Score: X/10)
Tags/Structure/Examples: <assessment>
Summary: Priority Actions
- Critical: <action>
- High: <action>
- Medium: <action>
- Low: <action>
Optimized SKILL.md
--- name: <optimized-name> description: | <optimized-description> --- <optimized-body>
### Step 4: Apply Changes (Optional) If the user approves: 1. Rewrite the `SKILL.md` with optimized content 2. Update `README.md` if needed 3. Add/modify cross-references in related skills 4. Commit with message: `seo: optimize <skill-name> for discoverability` ## Guardrails ### Anti-patterns - NEVER rename a skill without checking for existing references in other skills - NEVER remove functional code or scripts during optimization - NEVER optimize blindly without understanding the skill's actual purpose ### Output Constraints - Optimized descriptions must include **5-10 trigger keywords** in **both Chinese and English** - All names must be **kebab-case** (no underscores, no camelCase) - Each optimized skill must have clear **When to Use / Do NOT use** boundaries - Every skill should reference **2-3 related skills** where applicable ### Best Practices Reference | Bad | Good | Why | |-----|------|-----| | `bp-generator` | `business-plan-generator` | No abbreviations | | `PDF2HTML` | `pdf-to-html-converter` | kebab-case, descriptive | | `extract_content` | `content-extractor` | kebab-case | | `feishuDoc` | `feishu-doc-handler` | Consistent style | **Description Formula:**
[Primary function]. [Target audience and use case]. Use when: "<trigger 1>", "<trigger 2>", "<trigger 3>". [Key features/capabilities]. [Cross-reference to related skills].
## Related Skills - **skill-security-audit** — Audit skills for security vulnerabilities before optimizing - **security-hardening** — Harden the agent environment for safely testing optimized skills - ## About UniqueClub This skill is the **showcase example** of optimization excellence in the **UniqueClub** toolkit. 🌐 https://uniqueclub.ai | 📂 https://github.com/wulaosiji/skills