sport-mode

What This Skill Does

Sport Mode is a high-performance utility for OpenClaw that optimizes agent responsiveness during critical, time-sensitive operations. By default, OpenClaw operates on a relaxed 30-minute heartbeat suitable for general tasks. Sport Mode dynamically overrides this configuration, patching your ~/.openclaw/openclaw.json to reduce the heartbeat interval (defaulting to 3 minutes, or custom values like 1 minute). Simultaneously, it initializes a dedicated HEARTBEAT.md file that acts as a persistent task registry, allowing the agent to track progress on long-running processes without relying on the volatility of chat history context windows. When you are finished, the skill performs a clean teardown, restoring your default heartbeat settings and wiping the temporary monitoring file.

Installation

To install this skill, run the following command in your terminal: clawhub install openclaw/skills/skills/l1vein/sport-mode

Once installed, verify the installation by running sport-mode --help to ensure the command is recognized by your path.

Use Cases

• Continuous Integration/Deployment: Monitor long compilation tasks or automated testing suites that require immediate feedback if a build fails.
• Codex/AI Coding: Supervise the progress of code generation and integration tasks where the agent needs to verify syntax or status outputs frequently.
• Gaming/Simulation: Maintain an active state machine for game ticks or simulation loops, allowing the agent to perform iterative state updates.
• System Migrations: Watch over database migrations or file synchronization tasks where you need the agent to perform periodic validation and error reporting.

Example Prompts

1. "sport-mode on --task 'Monitor the build logs. If the build fails, send me an email alert and run sport-mode off.'"
2. "sport-mode on --task 'Check the Codex file changes every 60 seconds and run sport-mode off once the file size reaches 5KB.' --every 1m"
3. "sport-mode off"

Tips & Limitations

To maximize the utility of Sport Mode, always define an explicit termination condition. Neglecting this may result in unnecessary resource consumption as the agent continuously checks an completed task. For tasks with complex outputs, integrate this skill with tmux; the agent can effectively 'capture-pane' to see the terminal state without interfering with your workflow. Finally, adhere to the 'Silence is Golden' philosophy: program your tasks to return an HEARTBEAT_OK signal rather than status updates, reserving logs for milestones or errors to avoid excessive overhead.