The Agent Skills Directory

[COMMAND_EXECUTION]: The skill is instructed to execute the git log command via the History Scanner (as documented in references/intent-discovery.md) to extract commit messages that explain the rationale behind code changes. This is a functional requirement for the context-aware audit process.
[PROMPT_INJECTION]: The skill exhibits an indirect prompt injection surface as it is designed to ingest and interpret instructions and rationale from untrusted source material including project documentation (e.g., CLAUDE.md, ARCHITECTURE.md), source code comments, and git history.
Ingestion points: Documentation files, source code rationale markers (e.g., // RATIONALE), and git commit history.
Boundary markers: No specific boundary markers or instructions to ignore embedded agent directives are defined for the scanning process.
Capability inventory: The skill has the capability to read and write files within the project root and execute local git commands.
Sanitization: No sanitization or escaping of the extracted content is performed before it is processed by the intent discovery subagents.

code-audit