prd-task
Pass
Audited by Gen Agent Trust Hub on Mar 31, 2026
Risk Level: SAFEPROMPT_INJECTION
Full Analysis
- [PROMPT_INJECTION]: The skill processes external markdown files and extracts 'Verification' steps that are used as instructions for the agent's future tasks. This enables an indirect prompt injection attack where a malicious PRD file can embed instructions to manipulate the agent's behavior during the task execution phase.
- Ingestion points: Processes markdown documents found in the local environment (e.g.,
prd-*.md). - Boundary markers: No explicit markers are used to identify the PRD content as data rather than instructions.
- Capability inventory: The expected agent workflow includes code modification and environment verification, which provides a high-impact surface for successful injection attacks.
- Sanitization: The skill does not perform any validation or sanitization on the content of the markdown files before including it in the generated schema.
Audit Metadata