The Agent Skills Directory

[COMMAND_EXECUTION]: The skill executes shell commands to maintain project structure, specifically using mkdir -p to create goal-specific directories. It also invokes a CLI tool named plannotator to validate and 'gate' the generated facts.md and plan.md files, ensuring they meet specific criteria before proceeding to the next phase.
[DATA_EXPOSURE]: The skill instructions direct the agent to 'explore the codebase' to ground its understanding and validate implementation paths. This involves reading local files to extract facts and identify involved systems, which is the primary intended function of the skill.
[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data from user interviews and codebase files to generate goal outputs. While it lacks explicit boundary markers or sanitization for instructions that might be embedded in the analyzed code, the risk is addressed by the structured, multi-phase interview and gating process where the human user reviews the facts and plans.

plannotator-setup-goal