mj-01mvp-art

Pass

Audited by Gen Agent Trust Hub on Jun 30, 2026

Risk Level: SAFE
Full Analysis
  • [COMMAND_EXECUTION]: The skill utilizes local shell scripts (prepare-social-image.sh and upload-r2-asset.sh) to automate image processing and cloud storage operations. These scripts are implemented with defensive practices, including the use of set -euo pipefail and proper argument quoting to prevent shell injection during file handling.
  • [EXTERNAL_DOWNLOADS]: The skill depends on established third-party utilities, specifically ImageMagick and Cloudflare Wrangler, which are expected to be available on the execution host. It does not perform any unverified remote script downloads or arbitrary execution at runtime.
  • [CREDENTIALS_UNSAFE]: The skill adheres to high security standards for secret management. It explicitly instructs the agent to avoid logging or persisting API keys and relies on Wrangler's pre-existing shell authentication for its remote R2 storage operations.
  • [DATA_EXFILTRATION]: Network activity is restricted to the intended business purpose of uploading generated image assets to a Cloudflare R2 bucket. The skill provides clear guidelines for using verified public URLs and avoids any unauthorized data transmission.
Audit Metadata
Risk Level
SAFE
Analyzed
Jun 30, 2026, 08:22 AM
Security Audit — agent-trust-hub — mj-01mvp-art