The Agent Skills Directory

[SAFE]: The skill provides structured instructions and templates for security engineering tasks without executing any code or scripts. Access is restricted to benign file-reading tools (Read, Grep, Glob).
[PROMPT_INJECTION]: The skill includes a 'Prompt Injection Safety Notice'. This is a defensive mechanism designed to prevent the agent from being manipulated by untrusted user data, which is a security best practice for persona definitions.
[EXTERNAL_DOWNLOADS]: The skill references established security resources (OWASP, MITRE) for documentation and guidance. No executable code or unverified dependencies are fetched.

appsec-engineer