ql-intent-check
Pass
Audited by Gen Agent Trust Hub on Oct 8, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill's primary function is to ingest and compare external data sources, which creates a surface for indirect prompt injection. Malicious instructions embedded in the analyzed content (e.g., in commit messages or requirements documents) could attempt to influence the agent's behavior.
- Ingestion points: Processes
quantum.json(userIntent), design documents (docs/plans/*.md), project requirements (tasks/prd-*.md), and implementation artifacts (commit messages, code comments, and test names). - Boundary markers: The instructions do not define specific delimiters or security guardrails to distinguish between trusted audit instructions and untrusted data content.
- Capability inventory: The agent uses
Read,Grep,Glob, andBashtools to access and analyze the file system and version control history. - Sanitization: The skill does not describe any validation or sanitization logic for the text content being audited.
Audit Metadata