A high-performance deep research integration that connects AI agents to the LangGraph-powered DeerFlow engine for complex, multi-step information gathering.
Install & Download
1. ClawHub CLI
The fastest way to install a skill directly from the registry.
npx clawhub@latest install deerflow
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 deerflow 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 DeerFlow?
DeerFlow is a sophisticated LangGraph-based deep research engine designed to chain web searching, reasoning, and synthesis into comprehensive, structured reports. This skill enables Openclaw agents to submit and monitor research tasks via a running DeerFlow API, allowing for asynchronous execution of heavy research workloads without blocking the primary agent interface. By leveraging this addition to your library of Openclaw Skills, you can offload complex reasoning tasks to a dedicated engine while maintaining a lightweight local footprint.
DeerFlow Use Cases
- Executing deep research queries about market trends, technical specifications, or competitive landscapes.
- Running asynchronous reasoning chains through a dedicated LangGraph orchestration engine.
- Generating structured, long-form synthesis reports from multiple web search results.
- Monitoring multi-step research progress using persistent thread and run identifiers.
How DeerFlow Works
- The user or an agent triggers the research task using the /deerflow command.
- The submission script sends a POST request to the DeerFlow LangGraph API containing the research query and model configuration.
- The engine initiates a multi-step workflow and returns a thread_id and run_id to the caller.
- The monitoring script polls the API endpoint to track the execution status from running to success.
- Upon completion, the final synthesized report is retrieved and presented within the agent workspace.
DeerFlow Setup
To use this integration with your Openclaw Skills, you must have a DeerFlow instance reachable via HTTP. If you are deploying a new instance:
git clone https://github.com/bytedance/deer-flow.git
cd deer-flow
cp .env.example .env
# Configure your LLM keys in .env
docker compose up -d deer-flow-gateway deer-flow-langgraph
Configure the skill by setting the following environment variables:
export DEERFLOW_URL="http://your-host:2024"
export DEERFLOW_MODEL="minimax-m2.7"
DeerFlow Data Schema & Taxonomy
The skill manages research state through the following metadata and data structures:
| Attribute | Type | Description |
|---|---|---|
| thread_id | String | The unique identifier for a specific research conversation thread. |
| run_id | String | The unique identifier for a specific execution instance of the research graph. |
| status | String | The lifecycle state of the task (e.g., success, error, in_progress). |
| report | Markdown | The final synthesized research output generated by the LangGraph engine. |
| recursion_limit | Integer | Defines the maximum depth of the reasoning chain. |
DeerFlow Advanced Features
- Enable thinking_enabled mode to activate extended chain-of-thought reasoning for complex queries.
- Support for API-only deployment, minimizing resource overhead to approximately 2 GB of RAM.
- Multi-model compatibility including support for MiniMax M2.7, M2.5, and Kimi providers.
- Automated tool injection allowing other Openclaw Skills to programmatically invoke research workflows.
- Configurable recursion limits to manage the depth and cost of research tasks.
SKILL.md
Loading
Related Openclaw Skills
Writing Credibility Auditor
A sophisticated AI auditor that scans text for logical fallacies, unsupported claims, weasel words, and misleading statistics to generate a structured credibility report.
Lead Guardian
Lead Guardian is an AI-driven automation skill designed to instantly qualify real estate leads and alert agents to high-intent opportunities.
Compliance Guardian — FinCEN Real Estate Reporting
A specialized compliance tool for real estate professionals to navigate FinCEN reporting requirements for non-financed property transfers.
Polymarket Micro Session Edge Trader
An automated trading agent designed to exploit predictable mean-reversion patterns in Polymarket 5-minute crypto markets during global session opens.
Distill Conversation (Alembic)
A specialized tool that extracts first-principles knowledge from AI dialogues and generates structured, Obsidian-compatible Markdown notes.
Polymarket Micro Coin Lag Trader
A specialized trading skill that exploits price discovery delays between Bitcoin and altcoins in Polymarket 5-minute interval markets.
Featured*
