skill-sandbox（技能沙箱）

Skill Sandbox

Sandboxed installation pipeline for ClawHub skills. Install → Stage → Scan → Promote or Quarantine.

Quick Start

# Install a skill (stages, scans, auto-promotes if clean)
bash {baseDir}/scripts/skill-sandbox.sh <skill-name>

# Install a specific version
bash {baseDir}/scripts/skill-sandbox.sh <skill-name> --version 1.2.0

# Force install (bypass VirusTotal flags from clawhub)
bash {baseDir}/scripts/skill-sandbox.sh <skill-name> --force

# Re-scan a staged skill
bash {baseDir}/scripts/skill-sandbox.sh <skill-name> --scan-only

# Promote a quarantined skill after manual review
bash {baseDir}/scripts/skill-sandbox.sh <skill-name> --promote

# List all quarantined skills
bash {baseDir}/scripts/skill-sandbox.sh --list-staged

How It Works

1. Stage — Skill is installed to skills/_staging/<name> (never directly to live)
2. Scan — 5-layer automated security scan runs:

- File inventory (hidden files, symlinks, binaries) - Code pattern analysis (eval, exec, network calls, secret access, obfuscation) - SKILL.md instruction review (dangerous agent directives) - Dependency check (package.json install scripts, known-risky deps) - Publisher verification (metadata, origin registry)

1. Verdict:

- ✅ PASS (0 findings) → auto-promoted to skills/ - ⚠️ WARN (warnings only) → quarantined, manual review recommended - ❌ FAIL (critical findings) → quarantined, deep audit required

Scan Details

Critical Findings (auto-quarantine)

• eval(), new Function() — dynamic code execution
• Symlinks — path traversal risk
• postinstall / preinstall scripts in package.json — npm supply chain vector
• Dangerous SKILL.md instructions (disable security, exfiltrate, reverse shells, chmod 777)

Warning Findings (review recommended)

• Network calls (fetch, curl, axios, http)
• Shell execution (child_process, exec, spawn, subprocess)
• Environment/secret access (process.env, API_KEY, TOKEN)
• Base64 encoding patterns (potential obfuscation)
• File system writes
• Hidden files (excluding .clawhub/)
• Non-text binary files

Integration with Agent Workflows

For teams using security auditor agents (like Sentinel), the recommended flow:

1. Run skill-sandbox.sh for the fast automated scan
2. If WARN or FAIL → spawn your security agent for a deep LLM-powered audit of the staged files
3. After agent clears it → skill-sandbox.sh <name> --promote

Directory Structure

skills/
├── _staging/          ← quarantine area (gitignored)
│   └── <skill>/       ← flagged skills live here until promoted
├── skill-sandbox/     ← this skill
│   ├── SKILL.md
│   └── scripts/
│       └── skill-sandbox.sh
└── <other-skills>/    ← promoted (live) skills

Notes

• The _staging/ directory should be added to .gitignore
• Clean skills auto-promote — no manual step needed for safe installs
• The script returns exit codes: 0 (pass/warn), 2 (fail) for CI integration
• All scan patterns are static regex — no network calls, no external dependencies