tavily-web
Warn
Audited by Gen Agent Trust Hub on Sep 29, 2026
Risk Level: MEDIUMEXTERNAL_DOWNLOADSINDIRECT_PROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The installation instructions direct users to run
npx skills add -g BenedictKing/tavily-web. This command downloads and installs skill code from a remote GitHub repository belonging to an unverified third-party account ('BenedictKing'), which is outside the official or author-trusted ecosystem. - [INDIRECT_PROMPT_INJECTION]: The skill is designed to search, extract, and crawl web content, creating a significant attack surface for indirect prompt injection if malicious instructions are present on the target websites.
- Ingestion points: External data is ingested from the web via search results, URL content extraction, and site crawling (SKILL.md).
- Boundary markers: No delimiters or instructions are specified to prevent the agent from following directives found within the external content.
- Capability inventory: The skill has the capability to perform network requests and process the resulting content into the agent's context.
- Sanitization: There is no mention of sanitization, validation, or safety filtering of the content retrieved from the web.
Audit Metadata