self-driven

What problem does it solve?

Enables a software agent to continue working on long-running, well-scoped tasks without human intervention by providing a repeatable decision flow, periodic reminders, and strict exit criteria so users can leave work running overnight or in the background.

Core Features & Use Cases

• Autonomous Continuation: Uses periodic cron reminders and a clear decision flow to keep progress on tasks when the user is absent.
• Resilience & Strategy Switching: Advises switching approaches after prolonged stalls, attempts recovery for stuck background tasks, and proceeds with fallback actions when blocked.
• Safety & Verification: Enforces risk mitigation rules (safe git practices, no destructive database or production changes, no secret exposure) and requires testing and review before completion.
• Use Case: Leave an agent to run overnight to iterate on failing integration tests, process large datasets, or incrementally polish a feature until acceptance criteria are met.

Quick Start

Ask the agent to run overnight with a clear goal and explicit acceptance criteria so it can continue improving work until the criteria are satisfied.