Skills
Navigate
skills/edgeimpulse/agent-tools/build-custom-learning-blocks/Gen Agent Trust Hub

build-custom-learning-blocks

Pass

Audited by Gen Agent Trust Hub on Aug 23, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill references official Edge Impulse documentation and GitHub repositories (github.com/edgeimpulse/*) for templates and examples. These are verified vendor resources.
  • [COMMAND_EXECUTION]: The skill utilizes the edge-impulse-blocks CLI tool and Docker for building, local testing, and deploying custom learning blocks, which is necessary for the skill's stated purpose.
  • [DATA_EXPOSURE]: The instructions identify .ei-block-config as a sensitive metadata file and provide appropriate security guidance, recommending it be added to .gitignore for public or shared repositories.
  • [PROMPT_INJECTION]: The skill documents a data ingestion surface where custom training containers process external features.
  • Ingestion points: Training scripts read X_split_train.npy and Y_split_train.npy from a data directory.
  • Boundary markers: Absent.
  • Capability inventory: Training containers have file-system write access to the --out-directory and shell execution capabilities via the Docker ENTRYPOINT.
  • Sanitization: Not applicable to numerical feature data.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 23, 2026, 02:38 PM