linkedin-ad-library-api

Pass

Audited by Gen Agent Trust Hub on Sep 3, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data from an external source (LinkedIn Ad Library via Scavio API) which could contain malicious instructions.
  • Ingestion points: Ad copy, headlines, and advertiser names returned from the /ads/search and /ads/detail endpoints.
  • Boundary markers: The skill includes an instruction to "Only return what the API returned," which acts as a basic guardrail, but lacks explicit structural delimiters (e.g., XML tags or JSON schemas) to separate untrusted data from the agent's instructions.
  • Capability inventory: The skill enables the agent to perform network requests to the api.scavio.dev endpoint and interpret the returned JSON payload.
  • Sanitization: No evidence of sanitization or filtering of the ad content is present before the data is processed by the agent.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 3, 2026, 12:23 PM
Security Audit — agent-trust-hub — linkedin-ad-library-api