serp-feature-sniper
Pass
Audited by Gen Agent Trust Hub on Mar 29, 2026
Risk Level: SAFEPROMPT_INJECTION
Full Analysis
- [PROMPT_INJECTION]: The skill is designed to ingest and analyze untrusted content from the web using search tools and page fetching. This creates an indirect prompt injection surface where instructions hidden on external websites could potentially influence the agent's analysis or the content it generates in its reports. Ingestion points: Competitive pages fetched via fetch_webpage in Phase 2. Boundary markers: The blueprint generation phase does not specify the use of delimiters or warnings to ignore embedded instructions. Capability inventory: The skill utilizes web_search, fetch_webpage, and local file-writing operations. Sanitization: There is no evidence of filtering or sanitizing the content retrieved from external sources before it is processed into blueprints.
Audit Metadata