The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and act upon instructions found in external project files such as SPEC.md, DESIGN.md, and CONTEXT.md. This creates a vulnerability surface where malicious instructions embedded in these documentation files could influence the agent's behavior.
Ingestion points: SKILL.md explicitly directs the agent to load context from docs/dev/[feature-name]/SPEC.md, docs/dev/[feature-name]/DESIGN.md, docs/dev/[feature-name]/CONTEXT.md, and .agent/rules/general-rules.md.
Boundary markers: There are no defined delimiters or specific instructions for the agent to distinguish between legitimate implementation details and potential malicious injections within these files.
Capability inventory: The skill possesses file system capabilities, including reading existing code and writing/creating new files and directories.
Sanitization: The skill does not include any verification, sanitization, or validation steps for the content parsed from the documentation files.

developer