ai-content-pipeline

Pass

Audited by Gen Agent Trust Hub on Sep 2, 2026

Risk Level: SAFE
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill references installation instructions for the belt CLI from the official GitHub repository of inference-sh. This is a standard practice for utilizing the platform's tooling.
  • [COMMAND_EXECUTION]: The documentation includes several bash command examples using the belt CLI (e.g., belt app run). These are intended for building content pipelines and are presented as educational examples for the user.
  • [INDIRECT_PROMPT_INJECTION]: The skill describes pipelines that ingest external data (such as search results or blog content) and process it through LLMs. While this creates a theoretical surface for indirect injection, the skill acts as a workflow orchestrator and the risk is inherent to the use of LLM tools in a pipeline context.
  • [SAFE]: All external domains (inference.sh) and GitHub repositories (inference-sh/skills) are consistent with the skill's purpose and the vendor's infrastructure. No malicious patterns or obfuscation were detected.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 2, 2026, 02:18 PM
Security Audit — agent-trust-hub — ai-content-pipeline