agent-autopilot

Pass

Audited by Gen Agent Trust Hub on Mar 13, 2026

Risk Level: SAFE
Full Analysis
  • [COMMAND_EXECUTION]: The skill includes an initialization script (init.sh) and templates for periodic tasks that execute local shell commands. These commands are used for workspace setup, file management, and task tracking via the todo.sh utility.
  • [PROMPT_INJECTION]: The provided templates instruct the agent to assume a highly autonomous persona, explicitly directing it to proceed with tasks without waiting for user confirmation or providing standard acknowledgments. This is a functional requirement of the autopilot feature.
  • [DATA_EXPOSURE]: The skill creates and maintains internal state files (report-state.json) and historical logs within a dedicated memory/ directory in the agent's local workspace.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 13, 2026, 07:20 AM