The Agent Skills Directory

[COMMAND_EXECUTION]: The skill performs file system operations to read the repository codebase and write updates to local documentation files such as CONTEXT.md, docs/REQUIREMENTS-AND-NOTES.md, and docs/DECISIONS.md. These actions are within the intended scope of the skill's purpose as a documentation and planning assistant.
[PROMPT_INJECTION]: The skill demonstrates an indirect prompt injection surface because it reads untrusted data from the codebase while the agent is 'exploring' files. This creates a potential vector where malicious instructions embedded in source code comments or documentation could influence the agent's grilling session. However, the skill's restricted capabilities (limiting writes to documentation files) and lack of network access mitigate the risk.
Ingestion points: Reads files from the local repository during codebase exploration.
Boundary markers: None identified in the prompt templates to distinguish between instructions and ingested code content.
Capability inventory: Broad read access to the codebase and write access to specific documentation files (CONTEXT.md, docs/REQUIREMENTS-AND-NOTES.md, docs/DECISIONS.md).
Sanitization: No explicit sanitization or filtering of ingested file content is performed.

grill-me