let-me-know

What This Skill Does

The let-me-know skill is a critical utility for OpenClaw AI agents designed to bridge the communication gap during long-running background tasks. It ensures that users are never left wondering about the status of time-consuming operations like dependency installation, large-scale data processing, or automated testing suites. By implementing a standardized pre-flight notification system combined with an intelligent heartbeat mechanism, this skill guarantees that the agent remains transparent and responsive. It supports both "in-place" loop-based monitoring for synchronous execution and asynchronous cron-based heartbeats for detached background tasks, ensuring the user is always informed of progress, failures, or completion.

Installation

To integrate this skill, run the following command in your terminal: clawhub install openclaw/skills/skills/fogyoy/let-me-know Ensure your agent environment has the necessary permissions to read/write state files if you intend to use the cron-based heartbeat persistence features.

Use Cases

• Software Development: Monitoring multi-stage build pipelines or dependency compilations that exceed 3 minutes.
• Data Engineering: Tracking the progress of large ETL jobs or CSV batch processing tasks.
• System Administration: Automating server updates, backups, or security scans that require intermittent verification.
• AI Research: Running batch inferencing or fine-tuning scripts where status updates are necessary to catch model convergence or failure states early.

Example Prompts

1. "I need to run the full dependency update for this repo, which usually takes about 10 minutes. Use the let-me-know skill to keep me updated every 2 minutes."
2. "Please initiate the data migration script. It's a big task, so please send a start message and provide status heartbeats until it finishes."
3. "Start the system cleanup script. Set the heartbeat interval to 3 minutes and make sure to notify me immediately if the disk permissions cause it to fail."

Tips & Limitations

• Prefer In-Place Heartbeats: Always use the local loop (原地心跳) when possible. It is less prone to scheduling errors and naturally cleans itself up upon task completion.
• Dynamic Reporting: Never rely on static templates. The heartbeat messages should extract real-time metrics from your current task logs or state objects.
• Cron Hygiene: If you must use cron-based heartbeats for detached tasks, always include a cleanup handler in the finally block. Failure to remove cron jobs will result in persistent, ghost notifications that annoy the user.
• Error Handling: If the gateway times out during removal, do not ignore the error. Retry twice before scheduling a recovery cleanup job.