The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest untrusted project data to clarify domain models and maintain durable project records, creating a potential attack surface.
Ingestion points: As specified in SKILL.md and the reference files, the skill reads "relevant code/tests/configuration/runtime behavior" and existing project documentation (.learnings, .nongoals, CONTEXT.md, and ADRs).
Boundary markers: The instructions lack explicit boundary markers or directives to ignore instructions that might be embedded within the project files being analyzed.
Capability inventory: The skill has the capability to write to and update durable project records, including ADRs, .learnings, .nongoals, and domain language files.
Sanitization: The skill provides logical sanitization by explicitly instructing the agent in references/learnings.md to keep "secrets, credentials, and personal data out" of the project records it maintains.

amose