Security Audits

Scanned against catalog commit f566b24. Every scanner runs with its LLM stages switched off, so the same content always produces the same verdict. A result that cannot be reproduced is not an audit.

A flag is not a warning about the skill. This is a catalog of security methodologies, so a skill about SSRF contains the word SSRF and a skill about privilege escalation describes privilege escalation. Content-pattern scanners flag that by construction — which is why ssrf-hunting draws the loudest verdict here while being exactly what it says it is. Read the finding, then read the skill.

Clean is not a guarantee. Every one of these tools says so itself. They detect known and probable patterns; none of them certifies that content is safe, and none can catch a technique nobody has written a rule for yet.

Read anything before you install it. Every skill is public at securityskills/skills, and every loop prints in full on its own page.