The Agent Skills Directory

[COMMAND_EXECUTION]: The skill executes node scripts/run-ast-analysis.mjs to perform deep code inspection of the local codebase.
[COMMAND_EXECUTION]: Progress monitoring instructions involve using system utilities such as cat, jq, and watch to read local state files.
[PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection (Category 8) because its core functionality involves ingesting and acting upon untrusted project data. 1. Ingestion points: All files in the current directory are read and analyzed during the initial phases of the workflow. 2. Boundary markers: No delimiters or explicit safety prompts are used to isolate instructions that might be embedded in the analyzed code. 3. Capability inventory: The skill has the capability to write files, modify the project structure, and execute local scripts. 4. Sanitization: There is no evidence of sanitization or validation of the analyzed content before it influences the agent's automated implementation steps.

cruise-control