godot-asset-generator

Pass

Audited by Gen Agent Trust Hub on Sep 5, 2026

Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill documentation provides instructions for running Deno scripts (e.g., scripts/generate-image.ts, scripts/batch-generate.ts) with permissions for environment variable access, network communication, and filesystem read/write. These permissions are necessary for its core functionality of calling AI image generation services (OpenAI, Replicate, fal.ai) and saving the resulting assets locally.
  • [INDIRECT_PROMPT_INJECTION]: The skill includes an attack surface where untrusted data could influence agent actions if the batch specification file is malicious.
  • Ingestion points: The scripts/batch-generate.ts script accepts a local JSON specification file via the --spec parameter which contains prompt data.
  • Boundary markers: The documentation does not specify the use of delimiters or 'ignore' instructions for the data interpolated into the API prompts.
  • Capability inventory: The skill utilizes scripts with --allow-net for external API communication and --allow-write for local file creation.
  • Sanitization: There is no documentation regarding the sanitization or validation of the prompt strings provided in the JSON specification file.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 5, 2026, 08:06 PM
Security Audit — agent-trust-hub — godot-asset-generator