Submit a Skill

Submit your AI agent skill for security verification. Free for open-source projects.

The repository must contain a SKILL.md file at the root or in a skills/ directory.

The skill identifier as defined in your SKILL.md frontmatter.

Get notified when your skill passes verification.

Three-Tier Verification

Tier 1 — Scanned~2 minutes

37 deterministic pattern checks run against your SKILL.md. Detects prompt injection, data exfiltration, filesystem abuse, and other security anti-patterns.

Tier 2 — Verified~5 minutes

LLM-powered intent analysis evaluates whether the skill does what it claims. Checks scope alignment, capability escalation, and behavioral consistency.

Tier 3 — Certified1-3 days

Manual security review by a human expert. Deep-dive into edge cases, supply chain concerns, and runtime behavior. Highest trust level, valid for 6 months.

What Happens After You Submit

1

Submission Received

Your repo URL is validated and queued for processing.

2

Tier 1 Scanning

Automated pattern checks run within minutes.

3

Tier 2 Analysis

LLM intent analysis evaluates security and scope.

4

Review & Publication

Skill is published to the registry with its certification badge.