The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data from a local repository file which could contain malicious instructions designed to influence the agent's behavior.
Ingestion points: The skill reads .architecture/members.yml in Step 2 of the process defined in SKILL.md.
Boundary markers: No boundary markers or "ignore embedded instructions" warnings are present to delimit the external file content from the agent's system instructions.
Capability inventory: The skill utilizes the Read tool. Furthermore, the instructions under the "Adding New Members" section direct the agent to automatically create and add specialists to the team, which implies file-write capabilities.
Sanitization: There is no evidence of sanitization, validation, or escaping of the content read from .architecture/members.yml before it is processed and displayed.

list-members