seo-sxo
Pass
Audited by Gen Agent Trust Hub on Aug 12, 2026
Risk Level: SAFEPROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
- [PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection. It extracts data from user-provided URLs and search engine results (SERP) to derive user stories, personas, and wireframes. Maliciously crafted content on analyzed pages or in meta-tags could attempt to influence the agent's scoring or recommendations.
- Ingestion points: Target URL content fetched via
scripts/fetch_page.py, Google SERP snippets, People Also Ask (PAA) questions, and ad copy themes. - Boundary markers: No explicit delimiters or instructions to ignore embedded commands are used when processing external content.
- Capability inventory: Executes local scripts (
fetch_page.py,parse_html.py), performs web searches, and integrates with DataForSEO tools. - Sanitization: The skill body claims
fetch_page.pyis "SSRF-safe," but specific implementation details are not provided. - [EXTERNAL_DOWNLOADS]: The skill automatically retrieves content from external web pages provided by the user. While this is the intended purpose of the tool, it constitutes an ingestion point for external, unvetted data into the agent's context.
Audit Metadata