vike-web-fetch
Pass
Audited by Gen Agent Trust Hub on Sep 25, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process content from untrusted external URLs.
- Ingestion points: The
vike web fetch <url>command retrieves data from any user-provided URL into the agent's context. - Boundary markers: There are no explicit instructions for the LLM to ignore instructions embedded in the fetched content, although the extraction process (Tavily) may strip some active content.
- Capability inventory: The skill is documented to pass the extracted content to the Cerebras Llama 3.3 70B model to answer user questions.
- Sanitization: No sanitization or filtering of the fetched content is described in the skill definition.
- [EXTERNAL_DOWNLOADS]: The skill uses the
vike web fetchcommand to retrieve content from external domains and relies on external services (Tavily and Cerebras) to process that data. These are documented architectural components of the tool.
Audit Metadata