Tavily CLI
Fail
Audited by Gen Agent Trust Hub on Aug 30, 2026
Risk Level: HIGHEXTERNAL_DOWNLOADSREMOTE_CODE_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill instructions include downloading an installation script from Tavily's official CLI domain (cli.tavily.com).\n- [REMOTE_CODE_EXECUTION]: The installation process involves piping a remote setup script directly into a shell (curl | bash).\n- [INDIRECT_PROMPT_INJECTION]: The skill processes content from external websites, which introduces an ingestion surface for untrusted instructions embedded in search or crawl results.\n
- Ingestion points: The search, extract, crawl, and research commands in SKILL.md fetch data from arbitrary external URLs.\n
- Boundary markers: Absent. The skill does not define explicit delimiters or instructions to ignore embedded commands in the ingested data.\n
- Capability inventory: The skill possesses capabilities to search, crawl, and extract structured data from web sources.\n
- Sanitization: Present. The description states that results are ranked, filtered, and compressed for LLM use, providing a layer of data pre-processing.
Recommendations
- HIGH: Downloads and executes remote code from: https://cli.tavily.com/install.sh - DO NOT USE without thorough review
Audit Metadata