product-challenger
Pass
Audited by Gen Agent Trust Hub on Feb 19, 2026
Risk Level: SAFEPROMPT_INJECTION
Full Analysis
- Indirect Prompt Injection (LOW): The skill's primary workflow involves reading and analyzing external requirements documents (PRDs) and codebase structures, which are untrusted data sources.
- Ingestion points: SKILL.md (Workflow Step 1 and Document Analysis section) specifies reading requirement docs, PRDs, and scanning code structures.
- Boundary markers: The skill lacks explicit delimiters or specific instructions to ignore or isolate embedded prompts within the documents being analyzed.
- Capability inventory: The skill utilizes file system read access and performs outbound web searches based on keywords extracted from the documents.
- Sanitization: There is no evidence of sanitization, filtering, or validation of document content before it is used to generate search queries or influence the agent's analytical output.
Audit Metadata