valyu
Pass
Audited by Gen Agent Trust Hub on Mar 29, 2026
Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill instructs the agent to use the 'orth' CLI tool to perform API operations such as search, content extraction, and deep research task management.
- [EXTERNAL_DOWNLOADS]: The skill interacts with the Valyu API endpoints via the 'orth' CLI to fetch search results and research data from external web sources and proprietary datasets.
- [PROMPT_INJECTION]: The skill exhibits an attack surface for indirect prompt injection by ingesting and processing untrusted content from the web and third-party files. 1. Ingestion points: Untrusted data enters via search results from the /v1/search and /v1/answer endpoints, content extraction from the /v1/contents endpoint, and file attachments or URLs provided in the Create Task functionality (SKILL.md). 2. Boundary markers: No explicit delimiters or instructions to ignore embedded commands are present when the agent processes the retrieved external content. 3. Capability inventory: The agent can execute shell commands via the 'orth' CLI tool to interact with the Valyu API and manage research tasks (SKILL.md). 4. Sanitization: No sanitization, validation, or filtering of the ingested external content is described.
Audit Metadata