create-project-context
Pass
Audited by Gen Agent Trust Hub on Oct 4, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill reads external product requirement documents (PRD), agent guidance files, and the wider codebase to infer project conventions and architecture. This represents an indirect prompt injection surface where instructions embedded in these documents could influence the generated project context or agent behavior.
- Ingestion points: The skill ingests data from
it_{iteration}_product-requirement-document.md,it_{iteration}_PRD.json,AGENTS.md, and the codebase (when running in--mode yolo). - Boundary markers: There are no explicit boundary markers or instructions to the agent to disregard embedded directives within the ingested source files.
- Capability inventory: The skill has the capability to write to the local filesystem, specifically targeting
.agents/PROJECT_CONTEXT.md,.agents/state.json, and.agents/PROJECT_CONTEXT_archive.md. - Sanitization: The instructions do not define any sanitization, validation, or escaping logic for the content processed from external project files.
Audit Metadata