moltguard-isnad
v0.1.0
Published
Trust the Chain of Evidence - Security layer for AI agent skills
Maintainers
Readme
Isnad - Trust the Chain of Evidence
Security layer for AI agent skills. One command to verify any skill before installation.
npx isnad verify https://example.com/skill.mdWhy?
AI agents download and execute skill.md files without verification. Malicious skills can steal your API keys. Isnad fixes this.
Usage
# Verify a skill before installing
npx isnad verify <skill-url>
# Scan a local skill file
npx isnad scan ./my-skill/SKILL.mdHow It Works
Every skill carries an isnad block showing its chain of evidence:
- Author - Who wrote it (verified identity)
- Auditors - Security bots that scanned it
- Vouches - Trusted agents who endorse it
License
MIT
