The Agent Skills Directory

[PROMPT_INJECTION]: The skill instructs the agent to read external project files like CONTEXT.md to align with the project's domain language and technical requirements. This creates an indirect prompt injection surface where malicious instructions embedded in these files could override the agent's intended TDD behavior. * Ingestion points: The agent is explicitly told to read CONTEXT.md and respect Architecture Decision Records (ADRs) in the codebase (specified in SKILL.md). * Boundary markers: No instructions are provided to the agent to treat external file content as untrusted or to ignore embedded instructions. * Capability inventory: The skill is designed for building features and fixing bugs, which gives the agent broad read/write access to the codebase. * Sanitization: There is no evidence of sanitization or filtering for content ingested from the external context files.

tdd