shape
Pass
Audited by Gen Agent Trust Hub on Sep 27, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill directs the agent to gather facts by inspecting the repository environment, including code paths, git state, and manifests, which exposes the agent to potentially malicious instructions embedded in the codebase.
- Ingestion points: Processes branch data, HEAD, working tree, and project manifests as primary sources of information (references/_shared/orchestration/reality-check.md).
- Boundary markers: The instructions do not provide explicit delimiters or safety instructions to distinguish between data and embedded commands when reading repository content.
- Capability inventory: The shaping workflow is intended to influence product behavior specifications, architecture decisions, and implementation planning (SKILL.md, references/_shared/artifacts/spec.md).
- Sanitization: No mechanisms for sanitizing or validating repository facts are described before the information is used to update durable project specifications.
Audit Metadata