The Agent Skills Directory

[COMMAND_EXECUTION]: The skill executes local Node.js scripts (.agents/scripts/precheck.mjs and .agents/scripts/cooking.mjs) to perform state verification and ensure the environment is correctly configured before archival.
[COMMAND_EXECUTION]: It invokes git-commit to automatically commit any remaining work in the repository once the temporary cooking directory has been deleted.
[INDIRECT_PROMPT_INJECTION]: The skill ingests status and documentation data from workspace files (e.g., spec.md, goal.md, and review files) to evaluate if the completion criteria are met.
Ingestion points: Markdown files within the .agents/cooking/<feature>/ directory.
Boundary markers: Not explicitly defined in the prompt text, though the logic is guided by structured script output.
Capability inventory: Local script execution via Node.js, directory deletion, and git operations.
Sanitization: Relies on the programmatic output of the cooking.mjs status command to drive archival decisions rather than raw text interpretation.

archive