The Agent Skills Directory

[COMMAND_EXECUTION]: The skill utilizes the Bash tool to execute test commands derived from the project's environment and configuration files.
Evidence: SKILL.md (Phase 6) and reference.md detail the execution of commands like npx vitest run, python -m pytest, go test, and cargo test to verify code changes.
[INDIRECT_PROMPT_INJECTION]: The skill reads and acts upon task definitions from external Markdown files, which serves as an ingestion point for potentially malicious instructions.
Ingestion points: Phase 1 of SKILL.md specifies reading task data from files in specs/ or tasks/ using the Read tool.
Boundary markers: Absent. The skill does not provide instructions to the agent to distinguish between external task data and system instructions, nor does it use specific delimiters to wrap untrusted content.
Capability inventory: The agent has access to Bash (shell execution), Write, and Edit (file system modification) tools, which could be abused if the agent follows malicious requirements embedded in task files.
Sanitization: Absent. The skill does not implement filtering, validation, or escaping of the content read from the task files before using it to plan and implement code changes.

implement-with-test