creative-qa
Pass
Audited by Gen Agent Trust Hub on Sep 15, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest human-provided feedback, reasons, and verdicts to generate a structured
qaReport. This untrusted input could contain malicious instructions meant to influence the agent's behavior during normalization or affect downstream systems that consume the report. - Ingestion points: Human reviewer verdict and reasons (SKILL.md).
- Boundary markers: Absent. The instructions do not define clear delimiters or instruct the agent to treat the input data as strictly non-executable text.
- Capability inventory: The skill mentions executing 'owned CLI or script commands' (SKILL.md) and handing off feedback to a 'rerunTarget', which could involve file operations or tool executions depending on the environment.
- Sanitization: Absent. The workflow focuses on normalization and categorization but does not include steps to sanitize or filter the content of the provided reasons before embedding them in the structured output.
Audit Metadata