Proactive Tasks

Proactive goal and task management system. Use when managing goals, breaking down projects into tasks, tracking progress, or working autonomously on objectives. Enables agents to work proactively during heartbeats, message humans with updates, and make progress without waiting for prompts.

This package appears internally consistent for a proactive task manager, but review a few operational details before installing: - File writes: the script will create/update data/tasks.json, memory/WAL-YYYY-MM-DD.log, memory/working-buffer.md, and SESSION-STATE.md in your workspace. Ensure those locations are acceptable (no sensitive secrets) and set file permissions/retention accordingly. - Autonomous behavior: SKILL.md recommends cron/heartbeat patterns and autonomous 'isolated agentTurn' runs that may message humans. Confirm you want the agent to run background jobs and send updates automatically; restrict or review messaging targets in your environment. - Configuration placeholders: docs reference a person name ('Imran') and example cron commands—replace placeholders with your own targets and verify the paths (/openclaw, openclaw-runner) match your runtime. - Origin and versions: repository/homepage are not authoritative here (source unknown, package.json/README versions vary). If you need a higher trust level, request the upstream source or authorship verification and review the full scripts before granting autonomous operation. If you allow autonomous invocation, run the code in a controlled environment first and inspect created files to confirm no unexpected data is being recorded or transmitted.

Type: OpenClaw Skill Name: proactive-tasks Version: 1.2.3 The OpenClaw skill bundle 'proactive-tasks' provides a robust task management system for AI agents. All file operations are local, confined to the skill's data directory or the OpenClaw workspace's memory/session state files. There are no external network calls, no attempts to access sensitive system files or environment variables, and no use of obfuscation or malicious execution techniques. The markdown instructions (`SKILL.md`, `HEARTBEAT-CONFIG.md`) guide the agent towards autonomous, proactive task management and data persistence, without any evidence of prompt injection designed to subvert the agent's behavior for malicious purposes. The described cron jobs are for legitimate autonomous operation of the task management system, not for stealthy persistence or privilege escalation.