technical-blog-writing

Pass

Audited by Gen Agent Trust Hub on Sep 2, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSREMOTE_CODE_EXECUTIONCOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill fetches installation guidance and tool definitions from the inference-sh GitHub repository.
  • [REMOTE_CODE_EXECUTION]: Provides instructions to run remote platform applications such as exa/search for automated research and infsh/html-to-image for visual asset creation.
  • [COMMAND_EXECUTION]: Demonstrates the execution of Python scripts via the belt CLI to generate data-driven charts using the matplotlib library.
  • [INDIRECT_PROMPT_INJECTION]: The skill possesses an attack surface for indirect prompt injection. 1. Ingestion points: Results from exa/search tool (SKILL.md). 2. Boundary markers: Absent in the provided templates. 3. Capability inventory: Python execution, external tool invocation, and network posting (SKILL.md). 4. Sanitization: Absent.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 2, 2026, 02:20 PM
Security Audit — agent-trust-hub — technical-blog-writing