Claude Code -- Treeship -- Treeship

Integrate Treeship with Claude Code via the official plugin (recommended), CLAUDE.md, or MCP server.

Claude Code has three integration paths. The plugin is the recommended one — it auto-records every session via hooks, no manual treeship session start to remember.

Method 1: Official plugin (recommended)

The Treeship Claude Code plugin ships every session as a portable, signed receipt with zero configuration. SessionStart / PostToolUse / SessionEnd hooks fire automatically; the bundled @treeship/mcp server captures any MCP-routed tool calls; built-in tools (Read, Write, Edit, Bash, Grep, Glob) are captured via the PostToolUse hook.

Install the marketplace + plugin

claude plugin marketplace add zerkerlabs/treeship
claude plugin install treeship@treeship

The plugin lands at ~/.claude/plugins/cache/treeship/treeship/<version>/ and auto-loads in every subsequent Claude Code session. No restart needed for a session you start after install; existing sessions pick it up next time you open them.

Make sure the project is Treeship-initialized

cd your-project
treeship init   # if you haven't already

The plugin's hooks are gated on ./.treeship/ existing in the project root. Without it, the hooks silently exit zero so they never break Claude Code in unrelated projects.

Start coding

Open Claude Code in the project. The SessionStart hook automatically opens a Treeship session named after the project + timestamp. Every Read, Write, Edit, Bash, etc. is appended to the session's event timeline. When the session ends, the SessionEnd hook seals the receipt and surfaces the shareable session report URL back into the conversation context.

Updating: when a new plugin version ships, run claude plugin marketplace update treeship && claude plugin update treeship. Restart any open Claude Code sessions to pick up the new version.

What the plugin ships

Component	What it does
`.mcp.json`	Mounts `@treeship/mcp` via `npx -y` so MCP tool calls get signed receipts
`hooks/hooks.json`	SessionStart (auto-start), SessionEnd (seal + report URL), PostToolUse (capture built-in tools)
`monitors/monitors.json`	Live receipt/event counter streamed into Claude's context
`skills/treeship-session`	How to close with a real headline before SessionEnd auto-headlines
`skills/treeship-verify`	How to verify a receipt URL someone shares with you
`skills/treeship-report`	How to publish + share a session report URL

Method 2: CLAUDE.md instruction file (no plugin)

For users who don't want to install the plugin but still want Claude Code to use Treeship intentionally:

Install Treeship CLI

curl -fsSL treeship.dev/setup | sh

The /setup one-liner installs the CLI, runs treeship init, and instruments any AI agents it detects.

Drop CLAUDE.md into your project

curl -o CLAUDE.md https://raw.githubusercontent.com/zerkerlabs/treeship/main/integrations/claude-code/CLAUDE.md

Claude Code reads CLAUDE.md from the project root automatically. The template instructs Claude Code to wrap every shell command with treeship wrap --, set model/cost env vars, and close sessions with meaningful summaries. It also includes the trust block (what @treeship/mcp captures, what it doesn't, when data leaves the machine) so Claude Code trusts the MCP server.

Method 3: MCP server only

If you don't want hooks but want MCP tool calls captured:

// ~/.claude/mcp.json
{
  "mcpServers": {
    "treeship": {
      "command": "npx",
      "args": ["-y", "@treeship/mcp"],
      "env": {
        "TREESHIP_ACTOR": "agent://claude-code",
        "TREESHIP_HUB_ENDPOINT": "https://api.treeship.dev"
      }
    }
  }
}

Restart Claude Code after adding the config. This catches MCP-routed tool calls (e.g. brave_search, notion) but not Claude Code's built-in tools (Read, Write, Edit, Bash) — those bypass MCP entirely. For full coverage, use Method 1.

Live monitoring

While Claude Code works, watch the session in real time:

treeship session status --watch

The plugin also streams live receipt/event counters into Claude's context via its background monitor, so the agent itself can see the receipt being built without polluting the chat.

What appears in the receipt

Section	Plugin (Method 1)	CLAUDE.md (Method 2)	MCP only (Method 3)
Built-in tool calls (Read, Write, Edit, Bash, etc.)	Captured via PostToolUse `[AUTO]`	Via `treeship wrap` for shell `[AUTO]`	Not captured
MCP tool calls	Captured via `@treeship/mcp` `[AUTO]`	Captured via `@treeship/mcp` `[AUTO]`	Captured `[AUTO]`
Session lifecycle	SessionStart/SessionEnd `[AUTO]`	Manual `session start/close` `[EXPLICIT]`	Manual `[EXPLICIT]`
Model/cost	Via env vars `[EXPLICIT]`	Via env vars `[EXPLICIT]`	Via env vars `[EXPLICIT]`
Narrative	Auto-headline + skill-driven manual close `[MIXED]`	Via `--headline/--summary/--review` `[EXPLICIT]`	Not captured

Claude Code's built-in tools (Read, Write, Edit, Bash) are internal to the runtime, not MCP tools. They bypass the MCP bridge entirely. The plugin's PostToolUse hook is what closes that gap — without it, only MCP-routed tool calls would show up in the receipt timeline.

Verifying a session report

The URL printed by treeship session report is the human-readable session report at treeship.dev/receipt/<id>. Anyone you share it with can verify the embedded receipt themselves:

treeship verify https://treeship.dev/receipt/<id>

Verification is pure WASM (@treeship/core-wasm), works entirely offline, doesn't phone home.

Claude Code

On this page