The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill ingests content from external project files which could contain malicious instructions designed to influence the agent's output.
Ingestion points: The workflow involves reading docs/requirement.md, CONTEXT.md, architecture decisions (ADRs), and existing source code to inform the specification process as described in SKILL.md.
Boundary markers: The instructions do not define specific delimiters or warnings for the agent to treat the content of these files as untrusted data rather than instructions.
Capability inventory: The skill is permitted to read project files and write Markdown specifications to the docs/specs/ directory. It does not utilize tools for network communication, system-level command execution, or dynamic code evaluation.
Sanitization: There is no evidence of sanitization or filtering applied to the data read from the project files before it is processed by the agent.

write-spec