The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses shell commands to identify task files and execute project-specific quality gates.
The agent executes ls -t docs/tasks/*_prd.md to find input documents.
It runs build and test tools including make check, npm test, npm run lint, uv run pytest, and uv run ruff check based on project configuration files.
[PROMPT_INJECTION]: The skill is vulnerable to indirect prompt injection through the PRD (Product Requirement Document) files it processes.
Ingestion points: The agent reads and follows instructions from files matching docs/tasks/*_prd.md.
Boundary markers: There are no explicit delimiters or instructions provided to the model to ignore potential malicious commands embedded within the PRD files.
Capability inventory: The skill has the authority to write files (code implementation), execute shell commands (quality gates), and perform Git operations (git commit).
Sanitization: No validation or sanitization of the PRD content is performed before the agent begins autonomous implementation.

coder