exa-search

Warn

Audited by Gen Agent Trust Hub on Apr 15, 2026

Risk Level: MEDIUMEXTERNAL_DOWNLOADSPROMPT_INJECTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill instructions direct users to install a package from an external repository (BenedictKing/exa-search) using the npx command. This source is not recognized as a trusted organization and exists outside the skill author's verified infrastructure, presenting a supply chain risk.
  • [PROMPT_INJECTION]: The skill's core functionality involves ingesting untrusted data from the web via semantic search results, which is a known vector for indirect prompt injection attacks.
  • Ingestion points: Semantic search results and similar content discovery data retrieved via the Exa API from various external websites.
  • Boundary markers: No explicit delimiters or instructions are provided to the agent to treat the retrieved search results as untrusted content or to isolate them from system instructions.
  • Capability inventory: The skill allows the agent to perform semantic searches and discovery of structured research data.
  • Sanitization: There is no evidence of sanitization, validation, or content filtering mentioned in the skill's documentation to mitigate the risk of malicious instructions within search results.
Audit Metadata
Risk Level
MEDIUM
Analyzed
Apr 15, 2026, 06:10 PM
Security Audit — agent-trust-hub — exa-search