The Agent Skills Directory

[COMMAND_EXECUTION]: The skill utilizes the codemap CLI for codebase indexing and analysis. All bash commands are for legitimate code analysis and management purposes.
[EXTERNAL_DOWNLOADS]: The skill refers to the @ulpi/codemap package on NPM. The instructions specifically mandate that the AI must not perform automatic installations, ensuring human oversight for external software acquisition.
[PROMPT_INJECTION]: The skill ingests project code through search and analysis tools, which constitutes an indirect prompt injection surface. The risk is minimized by the tool's intended use-case and the author's inclusion of safety-first installation protocols.
Ingestion points: codemap search, codemap read, and codemap summary process source code files.
Boundary markers: None defined to isolate ingested code from AI instructions.
Capability inventory: Shell execution, file system read/write, and potential network access for updates (SKILL.md).
Sanitization: Not explicitly implemented in the provided skill instructions.

codemap