The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses the Bash tool to perform local file system operations, including globbing for directory discovery and executing MD5 hashing to verify the integrity of brand identity files. These actions are confined to the local project workspace and are consistent with diagnostic purposes.\n- [DATA_EXPOSURE]: The skill accesses project-specific configuration and documentation files (e.g., config.json, STATE.md, BRIEF.md). It does not target sensitive system credentials, environment variables, or private user data outside the intended project scope.\n- [INDIRECT_PROMPT_INJECTION]:\n
Ingestion points: The skill reads user-controlled content from files such as BRIEF.md, config.json, and various project chunks (SKILL.md).\n
Boundary markers: No explicit delimiters or instructions to ignore instructions within ingested data are present.\n
Capability inventory: The skill has access to Bash, Read, Glob, and Grep tools (SKILL.md).\n
Sanitization: No input sanitization is implemented. However, the risk is mitigated by the skill's internal instructions to operate as a 'pure pattern matching' tool that does not perform agentic reasoning or interpretation of the ingested content.

doctor