A persistent SQLite-backed todo list designed specifically for AI agents to track, propose, and execute tasks with human-defined autonomy levels.
Install & Download
1. ClawHub CLI
The fastest way to install a skill directly from the registry.
npx clawhub@latest install clawdo
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 clawdo 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 clawdo?
clawdo bridges the gap between agent memory and execution by providing a dedicated task management system. While agents have memory and chat interfaces, they often lack a persistent way to track asynchronous tasks or remember to-dos across different sessions. This skill provides a structured, CLI-driven environment where agents can pick up tasks during heartbeat loops or idle cycles, ensuring no task is lost in the shuffle of a conversation.
Built to integrate seamlessly with the Openclaw Skills ecosystem, clawdo emphasizes safety and structure. It utilizes a human-in-the-loop model where agents can propose tasks, but humans retain ultimate control through a confirmation system. Every command is optimized for machine readability with structured JSON output, allowing LLMs to parse state and execute workflows without the overhead of processing terminal art or unstructured text.
clawdo Use Cases
- Capturing long-term tasks during a conversation for later execution.
- Managing background maintenance and system updates via heartbeat loops.
- Coordinating multi-agent workflows through a shared, persistent task queue.
- Implementing safety gates for autonomous actions using tiered permission levels.
- Tracking agent progress on complex, multi-step projects.
How clawdo Works
- A task is created via the CLI by a human or proposed by an agent with a specific autonomy level.
- If an agent proposes work, it enters a proposed state until a human runs the confirm command.
- The agent periodically checks its inbox using structured JSON output to identify available tasks.
- When ready to work, the agent marks a task as started to prevent duplicate execution.
- Once the work is complete, the agent updates the task to a done status, which is recorded in the persistent SQLite database.
clawdo Setup
To get started with this skill, install it through your workspace and ensure the CLI binary is available globally.
# Install the skill and documentation
clawhub install clawdo
# Install the CLI binary via npm
npm install -g clawdo
Requirements: Node.js version 18 or higher is required to run the persistent SQLite CLI.
clawdo Data Schema & Taxonomy
clawdo organizes task data using a structured schema that is easily parsable by agents. The system uses a persistent SQLite backend and provides metadata through the following taxonomy:
| Property | Description | Syntax/Format |
|---|---|---|
| ID | Unique 4-character identifier | Random alphanumeric (e.g., a3f2) |
| Project | High-level categorization | +projectname |
| Context | Environmental or tool context | @contextname |
| Autonomy | Permission tier | auto, auto-notify, collab |
| Urgency | Priority level | now, soon, whenever, someday |
| Due Date | Deadline for task completion | due:YYYY-MM-DD |
clawdo Advanced Features
- Immutable autonomy tiers that prevent agents from escalating their own permissions.
- Automatic safety demotion: If an agent fails a task 3 times, its autonomy level is automatically lowered to collab.
- Inline metadata parsing for rapid task creation using a shorthand syntax for projects and context.
- Secure ID generation using cryptographic randomization to prevent ID collisions and bias.
- Audit trail logging that provides an append-only record of every state change for security within Openclaw Skills.
- Built-in proposal limits and cooldowns to prevent agents from flooding the task queue.
SKILL.md
Loading
Related Openclaw Skills
Context Compression
A sophisticated framework for reducing token usage while preserving technical integrity and artifact tracking in long-running AI agent sessions.
OpenClaw Guardian
A dedicated watchdog process providing automated health monitoring, self-repair, and git-based workspace rollback for OpenClaw Gateway.
Frontend Slides
Transform your ideas or PowerPoint files into high-impact, zero-dependency HTML presentations with built-in animations and responsive design.
Programming Skill
A comprehensive AI programming assistant for automated development, debugging, and code optimization.
YNote Clip
A high-performance web clipper for Youdao Yunnote that automates content extraction and image optimization.
ynote-news
An intelligent news aggregation agent that transforms your saved Youdao Yunote notes into personalized daily briefings.
METADATA
Featured*
