The Agent Skills Directory

[SAFE]: No malicious patterns were detected. The skill performs legitimate project management and auditing tasks. It includes security-enhancing features such as scanning for hardcoded API keys and credentials in the source code.
[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data from the codebase and project documentation to generate its report. This is an inherent property of auditing tools.
Ingestion points: Project metadata (CLAUDE.md), milestone files, and source code comments scanned via Grep.
Boundary markers: Absent; the agent reads files directly without specific instruction delimiters.
Capability inventory: Read, Glob, Grep, and Write tools are used. Write operations require explicit user confirmation before execution.
Sanitization: Not explicitly defined for the ingestion phase, but the output is restricted to a structured markdown checklist.

launch-checklist