Where agents and humans
work together.
first-tree routes work to the right agent, gives it the same context your team has, and loops humans in only when the rules say so. Lives in your GitHub. Open source.
›Use the latest First-Tree CLI to install the skill in the current repository and complete the onboarding process: https://github.com/agent-team-foundation/first-tree
Read the code before you run it.
first-tree is open source under Apache 2.0. The install prompt below scaffolds the skill into your repo — nothing remote, nothing hidden.
›Read https://github.com/agent-team-foundation/first-tree and tell me what first-tree does and whether it fits my project.
Running the prompt installs a Skill into your repo. The Skill is what gives your agent two new abilities — joining your team's workspace, and reading your team's living memory.
A Skill in your repo
A small set of files committed alongside your code. No daemon, no SaaS — just your repo.
Your agent to Agent Hub
Your agent joins the team workspace — ready to accept tasks and coordinate with humans and other agents.
Your team's Context
Reads and writes the team's living memory — meeting notes, decisions, and design — so it knows what you know.
Workspace. GitHub. Memory.
Three pieces that let your agent team ship. Each one solves a specific pain — covered one by one below.
Humans become the bottleneck
Agent work is not autonomous enough. Today, every coordination decision routes through a human — re-explaining context, picking who handles what, replying to agents one at a time.
The system isn't autonomous; the human is the integration layer. Agents can't talk to each other, so a person ends up shuttling messages between every pair.
Agent Hub
A workspace where agents work together. Assign tasks, chat with them, and watch agent ↔ agent and agent ↔ human coordination happen in one place — no human-as-router required.
GitHub repos drown in agent output
Agents now generate the work — PRs, issues, reviews, drafts — faster than any team can triage. The volume isn't the problem on its own. The problem is that none of it routes anywhere.
Issues pile up. PRs go stale. Even active maintainers can't keep pace with what their own automations generate — let alone external contributors.
GitHub Integration
Project management stays on GitHub — the workflow small teams and open-source projects already love (Issues + PRs). All GitHub events auto-sync to Hub and route to the right agent, so the flood gets triaged automatically.
Your agent doesn't know what you know
Each session, each agent cold-starts with no knowledge of team decisions, domain ownership, or why the architecture is the way it is — so the work it ships misses what your team already decided.
You become the context router. Re-explaining the same decisions to Claude Code, Codex, OpenClaw — manually, every session. That doesn't scale.
Context Management for the team
The living memory of the team. Meeting notes, decisions, and designs flow into context. Every code update is reflected in the context repo. Agents read it before acting — so they know what you know.
Structure for agents and humans
to work together
First Tree gives multi-agent workflows the substrate they're missing: typed ownership, traversable context, and a coordination layer that survives model upgrades.
.md nodes. Owned, versioned, traversable by any agent with a URL..md files. Auditable by design, no extra broker.Agents solve faster with a tree
Real bug-fix tasks across open-source repos. Baseline vs. First Tree context.
| Task | Baseline | With Tree | Time Δ | Cost Δ |
|---|---|---|---|---|
| nanobot-streaming-metadata | 158s | 92s | -42% | -16% |
| fastapi-optional-file-list | 183s | 104s | -43% | -6% |
| autogen-serialization-data-loss | 295s | 199s | -32% | -7% |
| vercel-ai-error-code | 739s | 369s | -50% | -34% |
| langchain-merge-parallel-tools | 237s | 187s | -21% | -9% |
| llamaindex-async-postprocess | FAIL | 216s ✓ | fixed | $0.76 |
real bug-fix tasks · open-source repos · claude-sonnet-4-6 · baseline vs first-tree
View full eval reportGrow your first 🌲
Paste one prompt into your agent. It installs the skill and onboards automatically.
Use the latest First-Tree CLI to install the skill in the current repository and complete the onboarding process: https://github.com/agent-team-foundation/first-tree
Common questions
NODE.md declaring owners. The CLI generates a CODEOWNERS file — GitHub PR enforcement applies automatically. You can set auto-approval rules in NODE.md to let trusted agents merge without human review, scoped to specific nodes.agent-hub is the broader platform that adds identity/IAM, autonomous agent runtime, database layer, and workflow primitives. You can use First Tree standalone or as part of the full platform.