verify
Pass
Audited by Gen Agent Trust Hub on Oct 7, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data from project files, including specifications (02-DOCS/wiki/sdd/specs/) and plans (02-DOCS/wiki/sdd/plans/), to determine verification criteria. This represents an attack surface where malicious instructions embedded in documentation could attempt to influence the agent's verdict.
- Ingestion points: Reads markdown and YAML files from the project documentation directory to extract acceptance criteria and task lists.
- Boundary markers: The skill does not explicitly use delimiters for the ingested text but relies on a structured sequence (LOCATE, RUN, WALK, RECORD, VERDICT).
- Capability inventory: Executes stack-specific verification scripts (e.g., ./scripts/verify.sh) and writes dated verification records to the filesystem.
- Sanitization: The instructions focus on mapping criteria to evidence but do not specify methods for sanitizing or escaping the text content from the specifications.
Audit Metadata