wulaosiji 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
- 探查 (Probe) 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.
约束 (Constrain) Set audit boundaries and不可降级的交付标准：7维评分体系、kebab-case命名、5-10个触发词、清晰的使用边界。不降级交付物——若技能缺少关键章节，必须补齐而非跳过。
证据 (Evidence) 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?
执行 (Execute) Create a structured report with these sections:
## 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: Recommended Addition:
4. Cross-References (Score: X/15)
Recommended Cross-References:
<related-skill>—
5. Guardrails (Score: X/15)
Recommended Additions:
6. README Integration (Score: X/10)
Status: ✅ | ⚠️ | ❌
7. Technical SEO (Score: X/10)
Tags/Structure/Examples:
Summary: Priority Actions
- Critical:
- High:
- Medium:
- Low:
Optimized SKILL.md
---
name: <optimized-name>
description: |
<optimized-description>
---
<optimized-body>
5. **验证 (Verify)**
用不同于生成路径的方式回读优化后的 SKILL.md：重新运行7维审计，确认分数提升、所有缺失项已补齐、触发词数量在5-10范围内、交叉引用双向有效。
6. **交付 (Deliver)**
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`
5. Clean up any temporary audit files
## Output
A structured **Skill Audit Report** with: overall score (X/100), per-dimension scores and recommendations, an optimized `SKILL.md` frontmatter and body draft, and a prioritized action list (Critical/High/Medium/Low). If the user approves, the optimized `SKILL.md` is written directly to the target file.
## 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
- **pitch-deck-creator** — Example of a well-optimized skill following these standards
## About UniqueClub
This skill is the **showcase example** of optimization excellence in the **UniqueClub** toolkit.
🌐 https://uniqueclub.ai | 📂 https://github.com/wulaosiji/skills