Skip to content
Get packs

DeerFlow Agent Bridge

Lets an agent send questions and files to a running DeerFlow server and read back its answers

Source Download skill
Repository
bytedance/deer-flow
Stars
84,876
Licence
MIT
Files
3 files

What it does

This skill teaches an AI assistant how to talk to DeerFlow, a separate multi-agent research platform that someone must already be running on their own server or machine. Once installed, when you ask for a deep research task or mention DeerFlow, the assistant can send your message over HTTP, start a conversation thread, and return DeerFlow's reply. It can also check whether DeerFlow is online, list its models, skills and agents, upload documents for it to read, and view past threads and memory.

Great for

  • Developers who already run a DeerFlow instance and want to drive it from Claude Code or another agent
  • Researchers who want to delegate long multi-step research to DeerFlow and collect the answer in one place
  • Teams testing or scripting DeerFlow's HTTP API

Skip it if

  • People without a running DeerFlow server: the skill only connects to one and will simply report it is unreachable
  • Sensitive documents or private data sent to a server you do not control, since uploads and messages leave your machine for whatever host DEERFLOW_URL points to
  • Use cases needing the agent to act on its own without a DeerFlow install, because it does none of the research itself

Try asking

  • “Ask DeerFlow to research the current state of solid-state battery startups and summarise the findings”
  • “Check whether my local DeerFlow instance is running and list which models it has”
  • “Upload this PDF to DeerFlow and ask it to extract the key risks”
  • “Continue my last DeerFlow thread and ask a follow-up question about the pricing section”
Step-by-step guide: how to use DeerFlow Agent BridgeInstall, first run, examples, tips and FAQs

Install in 3 steps

  1. 1

    Download the skill

    The ZIP contains the 3 original files, the MIT licence and an attribution note.

    claude-to-deerflow.zip
  2. 2

    Add it to your agent

    Claude Code (all projects)

    mkdir -p ~/.claude/skills && unzip claude-to-deerflow.zip -d ~/.claude/skills/

    Claude Code / Agent SDK (this project only)

    mkdir -p .claude/skills && unzip claude-to-deerflow.zip -d .claude/skills/

    Claude apps (web/desktop): Settings → Capabilities → Skills → Upload skill, and choose the ZIP.

  3. 3

    Ask for the task

    No special command needed — the agent reads the skill's description and loads it when your request matches. Try one of the prompts above.

Get better results

  • Use flash mode for quick factual questions and pro or ultra mode for research that needs planning and sub-agents, since higher modes take longer and use more resources
  • Point DEERFLOW_URL at a server you trust and avoid uploading confidential files to shared or public instances
  • Keep the returned thread_id if you want to return to the conversation later

What's inside

  • scripts/chat.sh
  • scripts/status.sh
  • SKILL.md

Redistributed unmodified from bytedance/deer-flow under the MIT licence. The assessment was drafted with AI from the skill's files on 2026-10-29. Skills can instruct your agent to run code — read SKILL.md and any scripts before installing.