The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes potentially untrusted data from external sources cited in project briefs, which could lead to instructions being executed by the agent.
Ingestion points: The skill reads research/*.md and files referenced under foundation/briefs/ (SKILL.md).
Boundary markers: The skill acknowledges risk by explicitly reading external material under the dx-references topic untrusted-content (SKILL.md).
Capability inventory: The agent has capabilities to read project files, write new frame.md files, and invoke other skills like dx-domain and dx-references (SKILL.md).
Sanitization: There is no explicit sanitization or filtering logic mentioned for the content of external briefs before they are processed during the framing interview.

dx-frame