newsletter-curation

Pass

Audited by Gen Agent Trust Hub on Sep 2, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill references an external setup guide hosted at raw.githubusercontent.com/inference-sh/skills/refs/heads/main/cli-install.md to install the necessary CLI tools for skill operation.\n- [INDIRECT_PROMPT_INJECTION]: The curation workflow involves fetching live data from external search assistants such as tavily/search-assistant and exa/search, which introduces a surface for indirect prompt injection where untrusted data from the web could contain malicious instructions.\n
  • Ingestion points: External search results fetched via belt CLI apps (Tavily, Exa search) in SKILL.md.\n
  • Boundary markers: None present in the provided shell command examples or content templates.\n
  • Capability inventory: Shell execution access (Bash with belt CLI) and network interaction capabilities.\n
  • Sanitization: No explicit filtering, validation, or sanitization of the fetched external content is described in the curation instructions.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 2, 2026, 02:18 PM
Security Audit — agent-trust-hub — newsletter-curation