Agentpack MCP server

Agentpack

Repo-native task continuity for AI coding agents.

Coding agents forget. Agentpack gives them the task state they need to continue.

Here it is live. Session 1: Claude Code investigates a flaky test and records what it learns through the Agentpack MCP tools. Session 2, next day, empty context: the agent loads the task state and picks up exactly where the first session stopped — no re-investigation:

Prefer to poke at it by hand? The same flow driven from the CLI is in docs/DEMOS.md.

Every session ends the same way: the context window gets compacted, the chat closes, the task waits until tomorrow. The next session starts from zero — re-reading files, rediscovering decisions, retrying approaches that already failed.

Agentpack keeps a small, reviewable task ledger in .agentpack/ inside your repo. Connected agents record durable state as they work — the goal, decisions, dead ends, verification evidence, checkpoints — and the next session loads it back and continues. That next session can be the same agent after compaction, a different client, or you returning next week.

• Local-first. Plain files in your repo. No cloud, no telemetry, no network calls.
• Agent-oriented. A local MCP server plus generated project instructions (AGENTS.md, CLAUDE.md, Cursor rules) tell agents when to load and record state.
• Human-friendly. The same state is available through the CLI for inspection, debugging, and manual handoff.

Quick start

Requires Node.js >= 20.

npm install -g agentpack-cli

cd path/to/your/repo
agentpack init                     # once per repo
agentpack install claude --write   # per client: codex | claude | cursor | claude-desktop

Restart or reconnect the coding-agent client. From then on the agent loads Agentpack context at session start, records durable decisions, sources, and evidence while working, and checkpoints meaningful progress.

Run agentpack doctor to verify the setup, and agentpack resume --preset agent --query "" to inspect the task state yourself.

See docs/INTEGRATIONS.md for client-by-client setup, including what each installer writes and why.

Codex, Claude Code, and Cursor include an optional builder for implementation work. The main agent can use it when you ask or your instructions allow delegation, and explains what it will hand off before starting. Small tasks stay with the main agent.

How it works

Shortened. The full README is on GitHub.