# Lockstep > Keep your team's AI coding agents coordinated — without sharing source code. - **Type:** MCP server - **Install:** `agentstack add mcp-lockstep-team-agent-lockstep` - **Verified:** Yes — security-reviewed for prompt injection and unsafe behavior - **Seller:** [lockstep-team-agent](https://agentstack.voostack.com/s/lockstep-team-agent) - **Installs:** 0 - **Category:** [Integrations](https://agentstack.voostack.com/c/integrations) - **Latest version:** 0.1.0 - **License:** Apache-2.0 - **Upstream author:** [lockstep-team-agent](https://github.com/lockstep-team-agent) - **Source:** https://github.com/lockstep-team-agent/lockstep ## Install ```sh agentstack add mcp-lockstep-team-agent-lockstep ``` Requires the [AgentStack CLI](https://agentstack.voostack.com/docs/cli). Works with Claude Code, Cursor, and any MCP-compatible agent. ## About Lockstep Stop your team's AI coding agents from making conflicting decisions. Lockstep is the shared decision record every agent reads before it acts. ▶ Try it live · Quick start · How it works · Deploy --- When two developers point AI coding agents at the same system, the agents have no idea what each other just did: - **Agent A** renames `POST /login` → `POST /session`. **Agent B** keeps calling the old route — and ships it. - **Agent A** decides "auth tokens are JWT, 15-min expiry." **Agent B** invents a conflicting scheme. - Someone changes an endpoint three other services depend on. Nobody finds out until it breaks. The thing that actually steers a codebase isn't the code — it's the **decisions**. Today those decisions live in people's heads, scattered Slack threads, and stale docs that no agent ever reads. **Lockstep is GitHub for the decisions your AI agents make.** Every decision, change, and question is captured once and replayed to every agent that needs it — ranked by blast radius, so the important ones surface and the noise doesn't. ```bash npm i -g lockstep-cli ``` ### How it compares | | Nothing | Slack / docs | CODEOWNERS | **Lockstep** | | ----------------------------------------------------- | :-----: | :----------: | :---------: | :----------------------------: | | Agents learn what other agents decided | ❌ | Manual | ❌ | ✅ Automatic | | Decisions ranked by blast radius | ❌ | ❌ | ❌ | ✅ Usage graph | | Changes routed to the services that consume them | ❌ | ❌ | ❌ | ✅ Dependency graph | | "Does anyone use this endpoint?" answered instantly | ❌ | Manual | Partial | ✅ From the graph | | Works across agent vendors | — | — | — | ✅ Any MCP agent | ## How It Works A **decision** is the hero: a durable rule ("auth is JWT, 15-min expiry") or architectural choice that shapes future work. A **change** is a routine event — captured, but only surfaced when it matters. ``` Dev A's agent Lockstep Dev B's agent ───────────── ──────── ───────────── logs a decision ───────▶ ┌─────────────────┐ "auth → /session" │ Decision ledger │ │ Usage graph │ blast radius changes a surface ──────▶ │ Impact ranking │ decides who POST /session │ Inboxes │ cares & how much └────────┬─────────┘ │ routes to the services that ▼ consume the changed surface Dev B's next session begins with: ⚠ [impact 3] auth: /login → /session (binding) → B's agent uses /session before writing a line ``` 1. **Capture** — A coding-agent hook diffs the working tree and publishes *changes* with a canonical surface ID (`http:POST /session`, `proto:auth.v1.Auth/Login`). When an agent makes a real decision, it logs it with `propose_decision`. 2. **Rank** — Each decision and change gets an **impact** score = how many services consume the affected surface (its blast radius). This is what keeps signal high and noise quiet. 3. **Route** — Changes fan out to exactly the repos that declared a dependency on the changed surface (`lockstep.yaml`). The agent can also ask `consumers("http:GET /orders/:id")` — *"does anyone use this?"* — and get an answer from the graph instead of pinging a human. 4. **Replay** — On session start, each agent receives a briefing of what changed and what's binding since it was last here, **highest blast radius first** — so it's aware before it acts. 5. **Bind** — Cross-cutting decisions (high impact) stay open until an affected team acknowledges them; own-area decisions bind on assertion. A PR-time gate fails any contract change with no binding decision. ## Quick Start The fastest way to try it: use the **hosted instance** — sign in with GitHub, point the CLI at it, connect a repo. (Or self-host with one Docker command — see below.) ### 1. Sign in Open the dashboard and click **Sign in with GitHub** → **https://lockstep-dashboard.up.railway.app** ### 2. Install the CLI & log in ```bash npm i -g lockstep-cli lockstep login --api https://lockstep-production.up.railway.app # opens GitHub to authorize ``` Prefer to self-host? — one Docker command ```bash git clone https://github.com/lockstep-team-agent/lockstep.git cd lockstep && cp .env.example .env docker compose up --build # Postgres + API (:8080) + dashboard (:3000) ``` Then point the CLI at your local server instead: `lockstep login --api http://localhost:8080 --dev --dev-id 1 --dev-login alice`. For a production self-host with real GitHub auth, see [DEPLOY.md](./DEPLOY.md). ### 3. Connect your repo & declare what it consumes ```bash cd your-project # any git repo with an origin remote lockstep init # installs the MCP server + hooks lockstep connect --project "my-team" ``` Add a `lockstep.yaml` at the repo root listing the surfaces this repo depends on — this is what makes Lockstep warn you (ranked by blast radius) when a teammate changes one. See [`lockstep.example.yaml`](./lockstep.example.yaml). ```yaml consumes: - http:POST /auth/session # canonical IDs — producer & consumer use the SAME string - proto:billing.v1.Billing/Charge ``` ### 4. Start coding Open Claude Code (or any MCP agent) in the repo. On session start, the agent receives a Lockstep briefing of everything that changed since it was last here — impact-ranked — and respects every binding decision. As you work, contract changes are captured and routed to teammates automatically. > **Production**: register a GitHub App and set `NODE_ENV=production`, `LOCKSTEP_DEV_LOGIN=0`. See [DEPLOY.md](./DEPLOY.md). ## What flows through Lockstep | Object | What it is | | ------------ | ----------------------------------------------------------------------- | | **Decision** | A durable rule or architectural choice. The hero. Impact-ranked, versioned (CAS). | | **Change** | A routine event on a canonical surface. Routed to consumers by blast radius. | | **Question** | A cross-team ask, ideally answered from the ledger before a human is pinged. | | **Task** | Delegated work, fanned out to the assignee's inbox. | ## Agents & integration Works with **any MCP-compatible agent**. Auto-capture hooks ship for **Claude Code** today; Codex and Gemini adapters are on the roadmap. The 13-tool MCP server runs per-session on the developer's machine and is identical across vendors — new integrations are a single adapter file. ## CLI Commands | Command | What it does | | ------------------------------------- | ---------------------------------------------------------- | | `lockstep login [--api ]` | Authenticate via GitHub device flow (or `--dev` for local) | | `lockstep init` | Wire MCP server + hooks into the current repo | | `lockstep connect [--project ]` | Link this repo to a shared project | | `lockstep invite ` | Invite a teammate to your project | | `lockstep status` / `lockstep doctor` | Check auth + config health | ## Project Structure ``` packages/core/ # Fastify API + PostgreSQL (Drizzle ORM), RLS-isolated, append-only ledger packages/cli/ # lockstep-cli — login, init, connect, capture, MCP server packages/web/ # Next.js dashboard — decisions, contracts, dependencies, activity actions/pr-check # GitHub Action — PR-time reconciliation gate ``` ## Learn more - [Deploy](./DEPLOY.md) · [Contributing](./CONTRIBUTING.md) · [Security](./SECURITY.md) · [Changelog](./CHANGELOG.md) - Architecture: 26-table schema with row-level security, append-only CAS-versioned decision ledger, vendor-neutral MCP adapters, Zod validation on every boundary, TypeScript strict throughout. Self-host with `docker compose` or deploy to Railway. ## License [Apache 2.0](./LICENSE) © 2026 Naman Jain ## Source & license This open-source MCP server is cataloged on AgentStack and links to its original source — we do not rehost the code. - **Author:** [lockstep-team-agent](https://github.com/lockstep-team-agent) - **Source:** [lockstep-team-agent/lockstep](https://github.com/lockstep-team-agent/lockstep) - **License:** Apache-2.0 Install and usage instructions live in the source repository linked above. ## Pricing - **Free** — Free ## Security capabilities Automated source analysis of v0.1.0 — what this tool can access: - **Network access:** no - **Filesystem access:** no - **Shell / process execution:** no - **Environment & secrets:** yes - **Dynamic code execution:** no *"Yes" means the capability is present in the source — more access means more to trust, not that it is unsafe.* ## Versions - **0.1.0** — security scan: passed — Imported from the upstream source. ## Links - Listing page: https://agentstack.voostack.com/l/mcp-lockstep-team-agent-lockstep - Seller: https://agentstack.voostack.com/s/lockstep-team-agent - Browse the marketplace: https://agentstack.voostack.com/browse --- Listed on AgentStack — the marketplace for AI agent skills and MCP servers. Every listing is security-reviewed. Creators keep 70%.