calibrate-lut

Pass

Audited by Gen Agent Trust Hub on Sep 8, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: Downloads the SAM3 segmentation model from Hugging Face, which is a well-known service for hosting AI models, using the hf CLI tool.\n- [COMMAND_EXECUTION]: Utilizes the calibrate-lut and calibrate-lut-classic command-line tools from the premiere-ai package, a resource owned by the skill author, to process calibration images.\n- [INDIRECT_PROMPT_INJECTION]: The skill is designed to process user-provided photos of calibration charts to build color-correction LUTs.\n
  • Ingestion points: Image file paths (e.g., chart_photo.png) provided as input to the calibration commands.\n
  • Boundary markers: None present.\n
  • Capability inventory: Shell command execution and file system writes for generating .cube files.\n
  • Sanitization: Input data is processed through the image-analysis logic of the premiere-ai package.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 8, 2026, 02:10 PM
Security Audit — agent-trust-hub — calibrate-lut