og-image-design

Pass

Audited by Gen Agent Trust Hub on Sep 2, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
  • [EXTERNAL_DOWNLOADS]: Fetches installation guidelines from a public repository on GitHub (inference-sh/skills).- [INDIRECT_PROMPT_INJECTION]: The skill provides an attack surface where untrusted data, such as blog titles or subtitles provided by a user, is ingested and processed by external tools.
  • Ingestion points: User-provided strings within the --input JSON for belt app run commands (SKILL.md).
  • Boundary markers: None explicitly defined in the prompt templates to distinguish between instructions and data.
  • Capability inventory: Execution of belt CLI commands via the Bash tool to interface with remote rendering and AI inference services.
  • Sanitization: None observed in the skill templates; sanitization is delegated to the downstream application logic.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 2, 2026, 02:18 PM
Security Audit — agent-trust-hub — og-image-design