skills/rheinmir/setup/fdk-uat/Gen Agent Trust Hub

fdk-uat

Pass

Audited by Gen Agent Trust Hub on Sep 21, 2026

Risk Level: SAFEREMOTE_CODE_EXECUTIONCOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTIONDYNAMIC_EXECUTIONEXTERNAL_DOWNLOADS
Full Analysis
  • [REMOTE_CODE_EXECUTION]: The skill downloads and executes shell scripts directly from a well-known service (raw.githubusercontent.com). This is used to verify that the installation process works for new users using the curl | bash pattern.
  • [EXTERNAL_DOWNLOADS]: The skill fetches bootstrap scripts, policy files, and configurations from external GitHub repositories.
  • [COMMAND_EXECUTION]: The skill performs significant repository operations, including git push --force-with-lease, git reset --hard, and git revert. It also utilizes the orca CLI to add repositories and create worktrees.
  • [DYNAMIC_EXECUTION]: The skill uses python3 -c to dynamically parse and process JSON output from CLI commands to extract identifiers for further operations.
  • [INDIRECT_PROMPT_INJECTION]: The skill executes user-provided checklist items as shell commands to verify that new capabilities are correctly delivered to the installation environment.
  • Ingestion points: User-supplied "checklist năng lực mới" (checklist of new capabilities) where each entry is treated as a command.
  • Boundary markers: None present to isolate or warn against instructions embedded in the user-provided checklist.
  • Capability inventory: The skill has access to shell execution, full Git repository control, and system environment management via orca.
  • Sanitization: No sanitization or validation is performed on the user-provided checklist items before they are executed.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 21, 2026, 10:52 AM