中文 EN

Scan Report

Scan New Files

✓ Trusted — Risk Score 5/100
Last scan:1 day ago Rescan
5 /100
plan
Think-first execution with approval gating. Use when work is complex, ambiguous, irreversible, multi-step, worth comparing before choosing, interrupted and needs recovery, or long-running enough to need a living plan.
Pure documentation skill providing a read-only approval-gated planning framework with no executable code or sensitive operations.
Skill Nameplan
Duration29.9s
Enginepi
✓
Safe to install
This skill is safe to use. It contains only documentation describing a planning methodology with no implementation code.

Findings 1 items

Severity Finding Location
Info
Toolbox audit concept implies capability not present Doc Mismatch
The SKILL.md mentions 'Check relevant installed skills' and 'Search skill registries (ClawHub, GitHub)' which implies potential network access for registry searching. However, no implementation code exists in the package.
Before presenting an execute-oriented plan:
- Check relevant installed skills first
- Check whether a ClawHub skill obviously fits
→ If registry searching is not implemented, remove references to ClawHub/GitHub to avoid confusion about actual capabilities.
 SKILL.md:241
ResourceDeclaredInferredStatusEvidence
Filesystem READ READ ✓ Aligned Documentation describes read-only exploration before approval
Network NONE NONE — No network operations in code (no code exists)
Shell NONE NONE — No shell execution described or implemented
Environment NONE NONE — No environment variable access described
Skill Invoke NONE NONE — Toolbox audit mentions checking skills but no invocation code exists
Clipboard NONE NONE — No clipboard access described
Browser NONE NONE — No browser access described
Database NONE NONE — No database access described

File Tree

2 files · 12.7 KB · 372 lines
Markdown 2f · 372L
├─ ▾ 📁 references
│ └─ 📝 patterns.md Markdown 84L · 4.6 KB
└─ 📝 SKILL.md Markdown 288L · 8.1 KB

Security Positives

✓ No executable code - pure documentation/methodology
✓ Explicit read-only stance until user approval
✓ Approval contract clearly defines when actions are permitted
✓ Surprise policy requires stopping and reporting unexpected changes
✓ No credential, file write, or network access operations
✓ No dependencies to audit for vulnerabilities
✓ Documentation is clear and matches the stated purpose