The Agent Skills Directory

[SAFE]: The skill follows established documentation practices and does not execute remote code, exfiltrate data, or attempt to persist across sessions. It uses standard tools like rg for codebase inspection and generates markdown artifacts in a local directory.
[INDIRECT_PROMPT_INJECTION]: The skill's primary function involves analyzing project source code, which is an external data source. This creates a surface for indirect prompt injection if the code being analyzed contains malicious instructions designed to influence the agent. The skill includes design features that mitigate this risk by instructing the agent to categorize information based on confidence levels (confirmed vs. inferred).
Ingestion points: Project source code is read using rg and rg --files as specified in the SKILL.md workflow.
Boundary markers: The skill does not explicitly define delimiters for code blocks in its instructions, although the provided templates use standard markdown formatting.
Capability inventory: The skill utilizes file system read capabilities via rg and file system write capabilities to the mydocs/codemap/ directory.
Sanitization: No specific sanitization or filtering logic for the analyzed source code content is described in the workflow.

codemap