skill-explorer

A systematic framework for discovering, evaluating, and comparing OpenClaw skills. Use when you need to find the right skill for a specific task, compare multiple options, or assess skill quality and safety.

Overview

This skill provides a battle-tested workflow for skill exploration:

1. Demand Analysis → Clarify what you need
2. Search & Discovery → Find relevant skills
3. Information Gathering → Collect metrics and details
4. Lateral Exploration → Find alternatives and competitors
5. Deep Dive → Download and inspect skill files
6. Safety Assessment → Security review
7. Synthesis & Report → Output comparison and recommendations
8. Decision Support → Present options for user approval

---

When to Use This Skill

Use skill-explorer when:

• You need to find a skill for a specific task
• You want to compare multiple similar skills
• You need to evaluate skill quality before installation
• You're doing security review of a skill
• You want to discover related or complementary skills

---

The Exploration Workflow

Phase 1: Demand Analysis

Clarify the requirements:

What task does the user need to accomplish?
What are the must-have vs nice-to-have features?
Any constraints (budget, platform, language)?
What's the priority: functionality, popularity, or safety?

Document the search criteria:

• Primary keywords
• Secondary/alternative terms
• Platform constraints (e.g., "must work with Chinese market")
• Safety requirements (e.g., "no network access")

---

Phase 2: Search & Discovery

Primary search:

clawhub search <primary-keyword>
clawhub search <related-term-1> <related-term-2>

Capture initial results: Record top 5-10 results with:

• Skill name
• Relevance score
• First impression of summary

---

Phase 3: Information Gathering

For each promising skill, collect:

clawhub inspect <skill-name> --json

Extract key metrics:

Metric	Why It Matters
stats.stars	Community quality indicator
stats.downloads	Popularity/visibility
stats.installsAllTime	Actual adoption
stats.installsCurrent	Active usage (low churn = good)
stats.versions	Maintenance activity
owner.handle	Developer reputation
updatedAt	Freshness

Read SKILL.md:

• Core functionality
• Use cases
• Prerequisites
• Integration requirements

---

Phase 4: Lateral Exploration

Find alternatives and complements:

# Search with related keywords
clawhub search <alternative-keyword-1>
clawhub search <alternative-keyword-2>

# Look for broader/narrower skills
clawhub search <broader-category>
clawhub search <specific-subtask>

Expand the candidate pool:

• Direct competitors (same function)
• Partial overlaps (related function)
• Complementary skills (work together)
• Domain-specific variants (e.g., "CN" versions)

---

Phase 5: Deep Dive

Download and inspect promising skills:

# Install to temp directory for inspection
cd /tmp && mkdir skill-check
cd skill-check && clawhub install <skill-name>

File structure analysis:

find . -type f | sort

Look for:

• SKILL.md - Main documentation
• scripts/ - Executable code
• package.json / requirements.txt - Dependencies
• _meta.json - Metadata
• Hidden files or unusual structures

---

Phase 6: Safety Assessment

Static code analysis:

# Check for suspicious patterns
grep -r -E "(eval\(|exec\(|Function\(|atob\(|btoa\()" .
grep -r -E "(crypto|encrypt|decrypt|base64|0x[0-9a-f]{20,})" .

# Check for network/API calls
grep -r -E "(https?://|api\.|token|key|secret|password)" .

# Check for file system operations
grep -r -E "(fs\.|writeFile|readFile|unlink|chmod)" .

Risk classification:

Level	Indicators
🟢 Low	Pure documentation, no scripts, no network
🟡 Medium	Scripts but official APIs only, read-only ops
🔴 High	Obfuscated code, unknown network endpoints, write operations

Special checks:

• VirusTotal warnings (verify if false positive)
• Hardcoded credentials or tokens
• External dependencies (npm/pip packages)
• Permission requirements (camera, microphone, etc.)

---

Phase 7: Synthesis & Report

Create comparison table:

| Skill | Stars | Installs | Safety | Best For |
|-------|-------|----------|--------|----------|
| skill-a | ⭐⭐ 25 | 45 | 🟢 | General use |
| skill-b | ⭐ 8 | 12 | 🟢 | Specific niche |
| skill-c | ⭐⭐⭐ 60 | 120 | 🟡 | Power users |

Write detailed analysis:

1. Executive Summary - Top 2-3 recommendations
2. Detailed Comparison - Feature-by-feature analysis
3. Safety Assessment - Security findings
4. Use Case Mapping - Which skill fits which scenario
5. Installation Guide - Prerequisites and steps

---

Phase 8: Decision Support

Present clear options:

## Recommendation

### Option A: Best Overall → [skill-name]
- Why: Highest quality + safety
- Trade-off: May lack advanced features

### Option B: Most Popular → [skill-name]
- Why: Community validated
- Trade-off: Higher complexity

### Option C: Specific Need → [skill-name]
- Why: Perfect fit for [specific use case]
- Trade-off: Less generalizable

Get user approval:

• Present options with trade-offs
• Ask for priority (functionality vs safety vs popularity)
• Wait for explicit approval before installation
• Document approved skills for installation queue

---

Best Practices

Do's

• ✅ Always check multiple similar skills (don't settle for first result)
• ✅ Download and inspect before recommending installation
• ✅ Document security findings clearly
• ✅ Consider the user's specific context (language, platform, expertise)
• ✅ Update EVOLUTION.md after installation

Don'ts

• ❌ Don't recommend based on downloads alone
• ❌ Don't skip security checks even for popular skills
• ❌ Don't install without explicit user approval
• ❌ Don't ignore VirusTotal warnings without investigation
• ❌ Don't overlook maintenance status (old updates = risk)

---

Integration with Evolution System

After completing an exploration:

1. Update EVOLUTION.md:

   • Add discovered skills to "待学习" list
   • Note priority and context
   • Link to exploration report

2. Create exploration report:

   • Save to reports/skill-exploration-{topic}.md
   • Include all findings and recommendations
   • Reference in future decisions

3. Track decisions:

   • Document which skills were approved/declined
   • Note reasons for future reference
   • Update after installation and usage

---

Example Output Template

# Skill Exploration Report: [Topic]

## Executive Summary
Recommended: [skill-name] (⭐⭐ 50 stars, 80 installs, 🟢 safe)

## Skills Discovered

### Tier 1: Strong Candidates
| Skill | Stars | Installs | Safety | Verdict |
|-------|-------|----------|--------|---------|
| ... | ... | ... | ... | ... |

### Tier 2: Alternatives
...

## Detailed Analysis

### [skill-name]
**Overview:** ...
**Strengths:** ...
**Weaknesses:** ...
**Safety Review:** ...
**Best For:** ...

## Recommendations

1. **Install:** [skill] - [reason]
2. **Consider:** [skill] - [reason]
3. **Skip:** [skill] - [reason]

## Next Steps
- [ ] User approval for recommended skills
- [ ] Security review for approved skills
- [ ] Installation and testing
- [ ] Update EVOLUTION.md

---

Related Skills

• marketing-mode - For go-to-market strategy after skill selection
• tweet-writer - For promoting discovered skills
• skill-creator - For creating your own skills based on gaps found

---

This skill helps you make informed decisions about skill adoption, ensuring you find the right tool for the job while maintaining security standards.