account-news-signals

Pass

Audited by Gen Agent Trust Hub on Jul 3, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSPROMPT_INJECTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill requires installing a vendor-managed package via npx skills add unifapi-agent/agents and connecting to an MCP server at https://mcp.unifapi.com. These are official resources provided by the author to support data retrieval.
  • [PROMPT_INJECTION]: The skill possesses an indirect prompt injection surface because it ingests untrusted data from public news articles and LinkedIn posts.
  • Ingestion points: The agent processes external content from news/search and linkedin/companies/{slug}/posts as instructed in SKILL.md.
  • Boundary markers: The instructions mandate that the agent cite specific sources and dates, providing a clear reference for the human operator.
  • Capability inventory: The skill is designed for read-only research and drafting; it cannot autonomously execute code, modify files, or perform network actions outside its specific research tools.
  • Sanitization: The skill implements a human-in-the-loop requirement, instructing the agent to flag single-source data for manual verification before outreach occurs.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 3, 2026, 04:06 PM
Security Audit — agent-trust-hub — account-news-signals