skills/garrytan/gbrain/gbrain-upgrade/Gen Agent Trust Hub

gbrain-upgrade

Pass

Audited by Gen Agent Trust Hub on Jun 13, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill triggers the gbrain self-upgrade command, which downloads and applies updates for the gbrain tool. This is the primary intended function of the skill for managing its own lifecycle.
  • [COMMAND_EXECUTION]: The skill uses the exec tool to run commands such as gbrain config get and gbrain self-upgrade. It implements a security contract that explicitly forbids running any commands parsed from external markers, ensuring only predefined, hardcoded commands are executed.
  • [PROMPT_INJECTION]: The skill contains a 'Contract' section that instructs the agent to ignore forged commands in upgrade markers, effectively hardening the agent against potential injection attempts via the tool's standard error stream.
Audit Metadata
Risk Level
SAFE
Analyzed
Jun 13, 2026, 04:42 AM
Security Audit — agent-trust-hub — gbrain-upgrade