The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill exhibits an indirect prompt injection attack surface as it is designed to ingest and process external project documentation that may be controlled by third parties.
Ingestion points: In SKILL.md, the skill identifies required inputs such as prior-year audit findings and engagement letter from auditor which are sourced from project owners or external workspaces.
Boundary markers: The skill lacks explicit delimiters or boundary markers to isolate the data content of these external files from the agent's instructions, potentially allowing embedded commands to influence the agent's output.
Capability inventory: The skill possesses read and search access to workspace files and generates governance artifacts (Attestation Prep Spec, Readiness Checklist) that impact project decision-making.
Sanitization: There is no evidence of input validation, filtering, or sanitization logic applied to the content of the external artifacts before they are interpolated into the generation process.

24-ai-agent-attestation-preparation-spec