← Back to Skills Marketplace
353
Downloads
1
Stars
1
Active Installs
1
Versions
Install in OpenClaw
/install smart-wake
Description
Prevent subagent timeout using checkpoint + cron wake + resume via session spawn mechanism.
Usage Guidance
This skill is instruction-only and internally consistent, but before installing check the following: (1) Confirm your platform's sessions_spawn/cron APIs are permissioned with least privilege and that the agent cannot create unrestricted system-wide cron jobs. (2) Define and audit where checkpoints and artifact paths are stored (memory/slot, file paths) and ensure sensitive data is not persisted in plaintext. (3) Ensure resume payloads are sanitized so secrets or unrelated context are not sent to spawned sessions. (4) Enforce retry_count and wake limits (to avoid accidental infinite wake loops) and test cancelation behavior. (5) If your environment requires an owner/authority review, note the SKILL.md owner line (platform-team) vs. registry owner ID—confirm this skill comes from an expected source before trusting it in production.
Capability Analysis
Type: OpenClaw Skill
Name: smart-wake
Version: 1.0.0
The 'smart-wake' skill is designed to manage long-running AI agent tasks by implementing checkpointing, scheduled cron wakes, and session resumption. The `SKILL.md` provides clear, structured instructions for the agent, focusing on state persistence, idempotency, and resource management within the OpenClaw platform. It explicitly limits tool usage to built-in mechanisms and includes operational rules to prevent infinite loops and ensure cleanup. There is no evidence of malicious intent, data exfiltration, unauthorized command execution, or prompt injection attempts to subvert the agent's core functions. The instructions are aligned with a legitimate operational goal for an AI agent.
Capability Assessment
Purpose & Capability
Name/description (prevent subagent timeout via checkpoint + cron wake + resume) match the SKILL.md instructions. The skill does not request unrelated binaries, env vars, or installs; required mechanisms referenced (sessions_spawn, cron, memory/slot) are consistent with the described functionality.
Instruction Scope
Instructions stay within the stated purpose (checkpointing, scheduling wake, spawning resume sessions). They do not ask the agent to read arbitrary system files or env vars. Caveats: storage location for checkpoints/artifacts is underspecified (mentions 'memory/slot' and artifact paths) which could lead to inconsistent implementations; resume payloads sent to sessions_spawn may inadvertently include sensitive context if not sanitized.
Install Mechanism
No install spec and no code files—lowest risk. The skill is instruction-only and will not write code or binaries to disk as part of install.
Credentials
The skill requests no environment variables or credentials, which is proportional. However it implicitly requires the agent/platform permission to register Gateway cron jobs and call sessions_spawn; ensure those privileges are scoped appropriately because cron/session capabilities can be powerful even without explicit credentials.
Persistence & Privilege
always is false and the skill does not request permanent presence or modify other skills. The workflow instructs creating and cancelling cron jobs for each task, which is reasonable for the stated purpose.
How to Use
- Make sure OpenClaw is installed (local or Docker)
- Run the install command in chat:
/install smart-wake - After installation, invoke the skill by name or use
/smart-wake - Provide required inputs per the skill's parameter spec and get structured output
Version History
v1.0.0
smart-wake` is a critical survival protocol designed to protect autonomous AI agents from being "killed" silently due to system timeouts when executing long, complex, or resource-heavy tasks (such as code compilation, large web scraping, or waiting for API responses). It transforms the agent from a passive listener into a proactive, self-sustaining worker.
**How to Use (For Agents):**
Whenever you are assigned a task that might exceed your given timeout limit, you must:
1. **Divide the work:** Break the task into smaller executable chunks (e.g., 5-minute blocks).
2. **Save your progress (Checkpoint):** Right before you suspect a timeout is approaching, save exactly what you have completed and what needs to be done next into a JSON checkpoint object.
3. **Set an Alarm (Cron Wake):** Use the `cron` tool to schedule a wake-up call for yourself in a few minutes.
4. **Pass the Baton:** In the cron payload, inject the `task_id` and the `next_actions`. When the cron wakes you (or spawns a fresh subagent session), you will read the checkpoint and resume exactly where you left off, completely bypassing the timeout limit!
Metadata
Frequently Asked Questions
What is Smart Wake?
Prevent subagent timeout using checkpoint + cron wake + resume via session spawn mechanism. It is an AI Agent Skill for Claude Code / OpenClaw, with 353 downloads so far.
How do I install Smart Wake?
Run "/install smart-wake" in the OpenClaw or Claude Code chat to install it in one step — no extra setup required.
Is Smart Wake free?
Yes, Smart Wake is completely free (open-source). You can download, install and use it at no cost.
Which platforms does Smart Wake support?
Smart Wake is cross-platform and runs anywhere OpenClaw / Claude Code is available (cross-platform).
Who created Smart Wake?
It is built and maintained by cong91 (@cong91); the current version is v1.0.0.
More Skills