An autonomous workflow skill that processes Todokan task boards by reading context, responding to comments, and managing task lifecycles through a review-loop.
Install & Download
1. ClawHub CLI
The fastest way to install a skill directly from the registry.
npx clawhub@latest install todokan-review-loop
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 todokan-review-loop 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 Todokan Review Loop?
The Todokan Review Loop is a sophisticated automation skill designed for Openclaw Skills that enables AI agents to act as Key Account Managers (KAM) within the Todokan project management ecosystem. It specifically targets tasks where AI is enabled and the status is set to doing. By analyzing the full context of a task—including its title, description, and chronological comment history—the agent provides high-quality, context-aware responses and ensures work packages are moved efficiently toward completion.
This skill is built to handle the complexities of human-AI collaboration in a Kanban environment. It doesn't just post updates; it reconstructs conversation timelines, performs internal research when needed, and adheres to a strict direct-answer-first policy. By integrating this within the broader Openclaw Skills framework, developers can ensure their AI agents provide professional, concise, and actionable contributions to team boards without manual intervention.
Todokan Review Loop Use Cases
- Autonomous polling of Todokan boards to handle tasks explicitly assigned to the AI.
- Streamlining the review process by moving completed work from doing to done (Review) with full execution logs.
- Managing complex project communication by acting as a single, consistent voice in task comments.
- Conducting internal research on specific tasks and summarizing findings directly within the board context.
How Todokan Review Loop Works
- The agent lists habitats and resolves the target board scope for the current execution cycle.
- It identifies tasks in the doing status that have the aiEnabled flag set to true.
- A temporal context gate is applied to reconstruct the comment thread chronologically, identifying the latest unresolved user questions.
- The agent performs a context sweep across related tasks and labels to ensure objective alignment.
- If the user request is complex, the skill spawns an internal research subagent to gather supporting data.
- A response is drafted following the conversation-aware reply policy, ensuring the first line addresses the user's primary intent.
- The agent posts the comment (or attaches a document for long responses) and transitions the task status to done (Review).
Todokan Review Loop Setup
To utilize this skill within Openclaw Skills, you must configure your environment with the necessary Todokan credentials.
# Set your Todokan API Key
export TODOKAN_API_KEY="your_api_key_here"
# Set the Todokan MCP Server URL
export TODOKAN_MCP_URL="https://your-mcp-endpoint.com"
Ensure that the tasks you want the agent to process have the AI Enabled toggle turned on in the Todokan UI.
Todokan Review Loop Data Schema & Taxonomy
The skill interacts with the Todokan data model to maintain consistency and traceability. Data is organized as follows:
| Entity | Data Point | Description |
|---|---|---|
| Task | Status | Moves tasks from doing to done (Review). |
| Task | Context | Combines Title and Description into a single semantic objective. |
| Comment | Threading | Reconstructs history using createdAt timestamps and unique IDs. |
| Document | Attachment | Created for responses exceeding 600 characters or containing complex code/logs. |
| Metadata | AI Flags | Only processes items where aiEnabled: true and assignee: 'ai'. |
Todokan Review Loop Advanced Features
- Temporal Context Gate: A mandatory preflight check that uses timestamps as first-class context to avoid answering outdated questions.
- Turn-Lock Mechanism: Enforces a deterministic selection of the active question to ensure replies are always relevant to the latest user input.
- Internal Delegation: Automatically spawns research subagents for factual verification or multi-step analysis while keeping the user-facing tone consistent.
- Objective Alignment Check: Verifies that every comment contains a clear Objective, Action, and Outcome before posting.
- Document vs Comment Policy: Dynamically switches between short chat-style updates and structured document attachments based on content length and complexity.
SKILL.md
Loading
Related Openclaw Skills
Fashion Colorize Shell
Converts garment sketches into high-fidelity, colorized shell jacket renders optimized for ecommerce product shots.
DoctorBot: HealthCheck (Free Edition)
A free diagnostic tool that performs non-invasive security and health audits for your Openclaw environment.
DoctorBot: CI Validator
A surgical validation tool for offline, deterministic checking of GitHub Actions, GitLab CI, and Keep workflows.
AppFlowy API Integration
A robust automation skill for managing AppFlowy Cloud workspaces, documents, and databases via the GoTrue API.
Todokan Kanban Task Manager
Todokan is a kanban-style task manager that enables AI agents to organize tasks, boards, and projects through the Model Context Protocol using Openclaw Skills.
Auto Researcher
An autonomous research assistant designed for deep investigations, data cross-verification, and professional citation report generation.
METADATA
Featured*
