z-market-research
Pass
Audited by Gen Agent Trust Hub on Aug 29, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill directs the agent to ingest content from untrusted external sources, including competitor websites, review platforms (G2, Capterra), and community forums (Reddit, Hacker News). This creates an attack surface for indirect prompt injection, where malicious actors could place hidden instructions on these pages to manipulate the agent's research findings or output.
- Ingestion points: Competitor discovery, pricing-page teardowns, and demand signal analysis involve fetching data from arbitrary third-party URLs in SKILL.md.
- Boundary markers: The instructions require labeling vendor copy as 'claims' rather than 'facts,' which provides a logical boundary for the agent's reasoning but does not provide technical sanitization.
- Capability inventory: The skill utilizes web retrieval tools to process external content.
- Sanitization: The skill lacks explicit instructions to sanitize or escape data retrieved from external sources before presenting it to the agent.
Audit Metadata