ShieldClaw Security Scanner

According to the Snyk ToxicSkills audit, 36% of ClawHub skills contain prompt injection. The existing free scanner misses 91% of threats. ShieldClaw was built to fix that.

ShieldClaw analyzes skills across 6 threat categories with 200+ detection patterns:

🧠 Prompt Injection Detection 🔑 Credential Access Detection 🌐 Suspicious Network Activity Detection 🔒 Code Obfuscation Detection 📁 Unauthorized File Access Detection 📦 Supply Chain Risk Detection

What you get: • shieldclaw.py — the scanner engine (zero dependencies, pure Python 3) • scan.sh — bash wrapper for quick scans • SKILL.md — install as an OpenClaw skill for conversational scanning • TUTORIAL.md — step-by-step setup and usage guide • Full documentation with batch scanning, JSON output, and CI/CD integration

QA tested: 51 tests passed across 6 phases. Zero false negatives. Zero false positives.