proactive-agent

To make AI agents proactive, you implement patterns like heartbeat-based self-improvement, reverse prompting, and autonomous crons. This shifts agents from passive task-followers to partners that anticipate needs and suggest tasks without external prompts.

Memory continuity uses a WAL Protocol, Working Buffer, and Compaction Recovery to preserve and recover context across sessions. This ensures long-running agents retain context and adapt to changing goals without losing prior state.

To set up a proactive agent, copy the provided assets to your workspace, answer the onboarding questions, and enable proactive mode. This activates the orchestration and memory protocols for continuous operation.

The best way to automate agent onboarding and growth is using built-in onboarding flows and growth loops. These sustain long-term usefulness by enabling heartbeat-based self-improvement and reverse prompting.

Agents lose context after compaction when recovery mechanisms are absent. Implementing Compaction Recovery alongside a WAL Protocol and Working Buffer ensures context remains preserved and recoverable during dynamic workflow operations.