The Agent Skills Directory

[SAFE]: The skill is a specialized security auditing tool that assists agents in identifying ReDoS vulnerabilities. Its instructions are purely educational and technical, focusing on legitimate security testing methodologies.
[COMMAND_EXECUTION]: The skill provides command templates using grep and semgrep to locate regex patterns in local source files. It also includes Node.js and Python script templates for performing timing measurements to confirm backtracking behavior. These are standard procedures for security auditing and do not involve remote code execution or data exfiltration.
[PROMPT_INJECTION]: The skill processes untrusted source code, creating a surface for indirect prompt injection.
Ingestion points: Local source files matching extensions defined in metadata.filePattern in SKILL.md.
Boundary markers: Not present; the skill treats file content as raw data for regex matching.
Capability inventory: Uses grep for searching and provides templates for local Node.js/Python script execution for timing tests as described in SKILL.md and references/poc-skeleton.md.
Sanitization: Not present; the skill relies on the agent's logic to extract and evaluate regex patterns. However, the instructions focus on structural analysis and timing verification rather than instruction execution.

redos