skills/eumetsat/skills/eumdac/Gen Agent Trust Hub

eumdac

Pass

Audited by Gen Agent Trust Hub on Nov 3, 2026

Risk Level: SAFECOMMAND_EXECUTIONDYNAMIC_EXECUTIONINDIRECT_PROMPT_INJECTIONDATA_EXFILTRATION
Full Analysis
  • [COMMAND_EXECUTION]: The skill manages EUMETSAT data via the eumdac CLI. It enforces execution safety by wrapping calls in timeout commands and requiring explicit user authorization for operations that involve significant data transfer or state changes.
  • [DYNAMIC_EXECUTION]: A Python script is dynamically generated via a shell heredoc in the SKILL.md rules to perform pre-download size calculations. This script is constructed from a local template and is used to provide the user with necessary information (product count and total size) before requesting download authorization.
  • [INDIRECT_PROMPT_INJECTION]: The skill processes external data in the form of satellite collection metadata and search results from EUMETSAT's APIs. Potential risks are mitigated by instructions that require the agent to present findings to the user and obtain permission before performing subsequent actions like downloads or tailoring.
  • [DATA_EXFILTRATION]: The skill interacts with sensitive local configuration files (e.g., ~/.eumdac/credentials) required for authentication. It mitigates exposure risks by explicitly forbidding the printing, logging, or committing of these credentials and by using placeholders in all instructional examples.
Audit Metadata
Risk Level
SAFE
Analyzed
Nov 3, 2026, 09:27 AM