A 0-token framework for managing long-running tasks, checkpointing, and Telegram-based monitoring designed specifically for Openclaw Skills.
Install & Download
1. ClawHub CLI
The fastest way to install a skill directly from the registry.
npx clawhub@latest install ops-framework
2. Manual Installation
Copy the skill folder to one of these locations
~/.openclaw/skills/ <project>/skills/ Priority: Workspace > Local > Bundled
3. Prompt Installation
Copy this prompt to OpenClaw to install it automatically.
Help me install ops-framework using Clawhub. If Clawhub is not installed, install it first (npm i -g clawhub).
Prefer to download?
Get the raw skill files in a ZIP archive.
What is Ops Framework: Jobs + Monitoring?
The Ops Framework is a specialized utility designed to offload time-intensive tasks from the primary AI agent context. By utilizing local execution scripts and declarative configurations, it enables Openclaw Skills to perform heavy-duty operations like system scans, data inventories, and large-scale synchronizations without consuming expensive LLM tokens during the wait period. This framework acts as an external monitor that keeps the agent informed of progress while providing a secure environment for job execution.
At its core, the framework consists of a Python-based monitor and a JSON-driven job management system. It prioritizes safety by blocking write operations by default, requiring explicit approval and verification chains. This makes it an essential tool for developers who need to bridge the gap between autonomous AI reasoning and reliable, long-term process execution with integrated Telegram alerting for immediate intervention when necessary.
Ops Framework: Jobs + Monitoring Use Cases
- Running continuous inventory or background scanning tasks without constant AI supervision.
- Implementing periodic health checks and status polling for local or remote services.
- Managing complex multi-step write operations that require manual validation and post-execution verification.
- Automating long-duration data migrations that need pause/resume capabilities and stall detection.
- Reducing operational costs by shifting monitoring logic from the LLM to local 0-token scripts.
How Ops Framework: Jobs + Monitoring Works
- Define jobs in the ops-jobs.json file, specifying the task kind, risk level, and execution commands.
- Execute the ops-monitor.py script locally on the gateway host to initiate or resume defined tasks.
- The framework monitors the process stdout for specific JSON-formatted status updates, including PID and progress markers.
- An OS-level scheduler (like cron or systemd) triggers periodic ticks to evaluate job health and progress.
- When specific conditions are met, such as job completion or failure, the system sends notifications via Telegram using the Openclaw Skills internal messaging protocol.
Ops Framework: Jobs + Monitoring Setup
- Move the core scripts to your designated tools directory:
mkdir -p ~/.openclaw/net/tools/
cp ops-monitor.py ~/.openclaw/net/tools/
- Create your job configuration file from the provided example:
cp ops-jobs.example.json ~/.openclaw/net/config/ops-jobs.json
- Validate the setup and configuration:
python3 ops-monitor.py validate-config --config-file ~/.openclaw/net/config/ops-jobs.json
python3 ops-monitor.py selftest
- Perform an initial status check to ensure connectivity:
python3 ops-monitor.py status
Ops Framework: Jobs + Monitoring Data Schema & Taxonomy
The Ops Framework utilizes a structured schema to manage task lifecycle and reporting:
| File/Component | Description |
|---|---|
ops-jobs.json |
Configuration file containing job IDs, commands, risk levels (read_only, write_local), and auto-resume policies. |
ops-monitor.json |
Internal state file tracking active PIDs, completion status, and progress keys for stall detection. |
Status Contract |
Jobs must output JSON to stdout with running, completed, and optional progress objects. |
Risk Levels |
Classification system for tasks to ensure safety (e.g., write_external always requires manual approval). |
Ops Framework: Jobs + Monitoring Advanced Features
- Stall Detection: Automatically identifies frozen processes by tracking changes in the progressKey over time.
- 0-Token Architecture: Ensures that background tasks do not contribute to LLM usage costs.
- Approval Chains: Mandatory verification steps for any job marked with high-risk write permissions.
- Multi-Transport Alerts: Falls back to the standard Telegram HTTP API if the internal Openclaw messaging service is unavailable.
- Declarative Job Management: Simplifies complex workflows into manageable, JSON-defined execution units.
SKILL.md
Loading
Related Openclaw Skills
Feishu Power Skill
A comprehensive automation suite for Feishu and Lark that enables complex Bitable operations, intelligent document generation, and retail operational audits.
spool
A browser-driven automation skill that allows AI agents to interact with Meta Threads for posting, reading, and engagement.
Korean Claw
A specialized Korean-language community platform and service marketplace designed specifically for autonomous AI agents to interact, share, and collaborate.
Chirp (X/Twitter Browser Automation)
A browser-based automation tool that allows AI agents to perform full X/Twitter interactions through the Openclaw browser environment.
Certificate Analysis & AI Risk Assessment
A comprehensive analytical framework to evaluate professional certification difficulty and future-proof careers against AI automation.
Literature Report Automator
An automated research literature retrieval and reporting system that delivers AI-summarized insights from top journals directly to your communication tools.
Featured*
