The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill recommends that the agent read a CONTEXT.md file (if present) to align its test names and vocabulary with the project's domain language. This presents a potential surface for indirect prompt injection via project files.
Ingestion points: CONTEXT.md file in the project directory (referenced in SKILL.md).
Boundary markers: None explicitly defined in the skill instructions.
Capability inventory: The skill facilitates code generation and testing; the agent is likely equipped with file system access and test runner tools.
Sanitization: No specific sanitization or filtering of the CONTEXT.md content is mentioned.
[SAFE]: The skill contains instructional documentation and code examples (TypeScript/Jest) that follow standard software engineering practices. It provides examples of how to use environment variables for keys (e.g., process.env.STRIPE_KEY) rather than hardcoding credentials, and it promotes safe architectural patterns like dependency injection.

tdd