venice-api-overview
Pass
Audited by Gen Agent Trust Hub on Sep 17, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill documents API endpoints that ingest and process external or untrusted data, creating a potential surface for indirect prompt injection.
- Ingestion points: The
POST /augment/scrapeandPOST /augment/searchendpoints described inSKILL.mdare designed to retrieve and process external web content into the agent context. - Boundary markers: The documentation does not specify the use of delimiters or 'ignore' instructions to isolate embedded commands in the content returned by these endpoints.
- Capability inventory: The skill identifies tools for web scraping, text parsing, and large language model inference via the
api.venice.aiinfrastructure. - Sanitization: No explicit sanitization, validation, or filtering mechanisms are described for the content ingested through the documented search and scraping utilities.
Audit Metadata