# Baton (npm · baton-run)

Run, hot-reload and screenshot Flutter, web, React Native and native iOS/Android apps.

- Trust score: 91/100 (high trust)
- Change this week: +3
- Registry status: active
- Liveness: live
- Owner verified: no
- Last scored: 2026-11-05

## Components

- npm · `baton-run`: 91/100 (this document), [markdown](https://verifymcp.io/servers/kanumuri9593-baton/baton-run.md), [page](https://verifymcp.io/servers/kanumuri9593-baton/baton-run)

## Channel facts

- Registry: `npm`
- Package: `baton-run`
- Version: `0.2.10`
- Transport: `stdio`

## Trust breakdown

How this component scores in each security and reliability category. Every signal is checked automatically from public evidence about the published package, including repeated runs of it in an isolated sandbox, and we only credit what we can confirm. Scores are 0–100 per category. Scoring method: https://verifymcp.io/docs/scoring (what has changed: https://verifymcp.io/docs/scoring/changelog)

Scored 2026-11-05.

- **Supply Chain Security**: 98/100
  - No malware found by supply-chain analysis.
  - No known CVEs affecting this package version or its production dependencies.
  - No install/post-install scripts declared.
  - 32 of 120 dependencies flagged as unhealthy (2 deprecated).
- **Provenance & Transparency**: 100/100
  - Source repository is publicly reachable at the declared URL.
  - Cryptographically verified build provenance (signed, bound to kanumuri9593/Baton).
  - Clear OSI-approved license (MIT).
  - Actively maintained (last published 24 days ago).
  - Publishes a security disclosure policy (SECURITY.md).
- **Schema Quality & AI Usability**: 79/100
  - 100% of prompts and resources have a non-trivial description (not blank, and not just the item's name).
  - AI-judged instruction clarity (good).
  - Tool/resource definitions use about 2810 tokens (~93/item across 30 items; 29 tools + 1 resources), lean.
  - Usage-examples check failed: none of the tools include examples.
- **Stability & Change Management**: 77/100
  - Stability observed for 23 of 30 days with no destabilising changes; credit accrues until the full window elapses.
- **Tool Coverage**: 86/100
  - 100% of tools have a non-trivial description (not blank, and not just the tool's name).
  - 58% of tool parameters carry a description.
- **Tool Safety**: 100/100
  - No prompt-injection markers were found in the server instructions, tool names or descriptions we captured.
  - All 2 tool(s) whose name or description implies an irreversible operation declare an MCP destructiveHint annotation.
  - An AI judge read all 30 captured unit(s) of tool text and found none that tries to manipulate the model reading it.
- **Capabilities**: 100/100
  - Implements a supported MCP spec version (2025-11-25); the latest is 2026-07-28.
  - Supports UI / widget rendering.

## Install

### How do I install the Baton MCP server?

Baton runs locally as an npm package, launched with npx -y baton-run. Ready-made configuration for Claude, Cursor, VS Code, Codex and 5 more is on this page, copied from each client's own documentation.

### Claude

```bash
claude mcp add kanumuri9593-baton -- npx -y baton-run
```

### Cursor

```json
{
  "mcpServers": {
    "kanumuri9593-baton": {
      "command": "npx",
      "args": [
        "-y",
        "baton-run"
      ]
    }
  }
}
```

### VS Code

```json
{
  "servers": {
    "kanumuri9593-baton": {
      "command": "npx",
      "args": [
        "-y",
        "baton-run"
      ]
    }
  }
}
```

### Codex

```bash
codex mcp add kanumuri9593-baton -- npx -y baton-run
```

### opencode

```json
{
  "$schema": "https://opencode.ai/config.json",
  "mcp": {
    "kanumuri9593-baton": {
      "type": "local",
      "command": [
        "npx",
        "-y",
        "baton-run"
      ],
      "enabled": true
    }
  }
}
```

### OpenClaw

```bash
openclaw mcp add kanumuri9593-baton --command npx --arg -y --arg baton-run
```

### Hermes

```yaml
mcp_servers:
  kanumuri9593-baton:
    command: "npx"
    args: ["-y", "baton-run"]
```

### Netclaw

```json
{
  "McpServers": {
    "kanumuri9593-baton": {
      "Transport": "stdio",
      "Command": "npx",
      "Arguments": [
        "-y",
        "baton-run"
      ]
    }
  }
}
```

### Vellum

```bash
assistant mcp add kanumuri9593-baton -t stdio -c npx -a -y baton-run
```

### Other

```json
{
  "mcpServers": {
    "kanumuri9593-baton": {
      "command": "npx",
      "args": [
        "-y",
        "baton-run"
      ]
    }
  }
}
```

## Changelog

Every change recorded for this component, newest first. Days that predate change tracking, or that we cannot explain, say so: "we were watching and nothing happened" and "we were not watching" are different claims.

### 2026-11-04 (score 91, +1)

No change was recorded against any check on this day. Stability & Change Management went from 70 to 73. That category is still filling its 30-day observation window: 21 days of observed history at the previous scan, 22 at this one. The score rises as the window fills, whether or not the server changes.

### 2026-11-02 (score 90, +1)

No change was recorded against any check on this day. Stability & Change Management went from 63 to 67. That category is still filling its 30-day observation window: 19 days of observed history at the previous scan, 20 at this one. The score rises as the window fills, whether or not the server changes.

### 2026-10-31 (score 89, +1)

No change was recorded against any check on this day. Stability & Change Management went from 57 to 60. That category is still filling its 30-day observation window: 17 days of observed history at the previous scan, 18 at this one. The score rises as the window fills, whether or not the server changes.

### 2026-10-28 (score 88, +1)

No change was recorded against any check on this day. Stability & Change Management went from 47 to 50. That category is still filling its 30-day observation window: 14 days of observed history at the previous scan, 15 at this one. The score rises as the window fills, whether or not the server changes.

### 2026-10-26 (score 87, +1)

No change was recorded against any check on this day. Stability & Change Management went from 40 to 43. That category is still filling its 30-day observation window: 12 days of observed history at the previous scan, 13 at this one. The score rises as the window fills, whether or not the server changes.

### 2026-10-24 (score 86, +1)

No change was recorded against any check on this day. Stability & Change Management went from 33 to 37. That category is still filling its 30-day observation window: 10 days of observed history at the previous scan, 11 at this one. The score rises as the window fills, whether or not the server changes.

### 2026-10-22 (score 85, +4)

- [functional improvement] Stability: unverified → 0.30

### 2026-10-13 (score 81, +31)

- [security improvement] Injection markers: unverified → pass
- [security] First check of Judged manipulation: pass
- [security] Stability: Stability not yet verified: not enough scan history yet (needs a 30-day window).
- [functional improvement] Tool coverage: unverified → 100
- [functional improvement] Schema quality: unverified → 100
- [functional improvement] MCP protocol: unverified → pass
- [functional] First check of Schema quality: fail
- [functional] First check of Capabilities: pass
- [functional] First check of Tool coverage: 58
- [functional] First check of Destructive annotations: 100
- [functional] First check of Schema quality: pass
- [functional] First check of Schema quality: good

## MCP tools (29)

### `diagnose` (~121 tokens)

Search bounded log and network evidence across projects. Defaults to errors only and 20 findings. Findings are grouped by session; repeats of the same error are counted once (×N), and a stack is collapsed to its first app frames. Pass a trace ID and errorsOnly=false to follow a request across Node services. Use exact session IDs to limit scope.

Input parameters:

- `errorsOnly` (boolean)
- `limit` (integer)
- `query` (string)
- `root` (string)
- `sessions` (array)
- `since` (number)

### `run_workflow` (~81 tokens)

Launch up to eight project targets in dependency order and wait for each to become ready, in one call. Stops launching dependents on failure. Returns compact session IDs, URLs and errors; retains sessions for debugging. Use browser/device tools next to validate the actual flow. Paths must be absolute.

Input parameters:

- `name` (string, required)
- `steps` (array, required)

### `inspect_project` (~60 tokens)

Read fresh launch sources, configuration problems, nested projects and guidance file paths. Does not run commands or expose environment values. Use before choosing an environment, checkout and device; review screenshot evidence rather than treating capture as visual approval.

Input parameters:

- `cwd` (string)

### `list_targets` (~44 tokens)

List everything runnable in a project: launch.json configs, package.json dev scripts, Flutter entrypoints.

Input parameters:

- `cwd` (string): Project directory. Defaults to the daemon working directory.

### `list_sessions` (~21 tokens)

List running sessions with their status, capabilities, URLs and device.

### `list_checkouts` (~48 tokens)

List git checkouts this project can run from: This checkout, linked worktrees, local branches, remotes.

Input parameters:

- `cwd` (string): Project directory. Defaults to the daemon working directory.

### `run_target` (~126 tokens)

Start a target by name. Returns the new session id to use with the other tools. Optional branch or checkout runs from a git worktree without moving the current folder.

Input parameters:

- `branch` (string): Git branch or remote ref to run from a Baton-owned copy. Mutually exclusive with checkout.
- `checkout` (string): Path to an existing worktree to attach. Mutually exclusive with branch.
- `cwd` (string)
- `deviceId` (string): Force a specific device or simulator.
- `target` (string, required): Target name or an unambiguous substring of it.

### `hot_reload` (~69 tokens)

Hot reload a session, keeping app state. Flutter only. Returns code 0 on success; a non-zero code carries the compile error.

Input parameters:

- `all` (boolean): Reload every running session.
- `reason` (string)
- `session` (string): Session id. Omit with all=true.

### `hot_restart` (~51 tokens)

Hot restart a session, dropping app state. For web dev servers and plain processes this reboots the process.

Input parameters:

- `all` (boolean)
- `reason` (string)
- `session` (string)

### `stop_session` (~34 tokens)

Stop one session, or every running session with all=true.

Input parameters:

- `all` (boolean)
- `session` (string)

### `forget_session` (~45 tokens)

Remove stopped sessions from the list. Running sessions must be stopped first. Use all=true to clear every stopped session.

Input parameters:

- `all` (boolean)
- `session` (string)

### `read_logs` (~137 tokens)

Read recent output from a session, newest last. Use filter to grep for an error: it searches all retained output and returns the last `tail` matching lines. Works for a currently running session id, and for a past run (its own session id, or the runId from list_run_history) even after it has stopped or the daemon has restarted.

Input parameters:

- `filter` (string): Case-insensitive regular expression.
- `session` (string, required): A live session id, or a past run's session id / runId.
- `tail` (number): How many lines (default 200); with a filter, how many matching lines.

### `list_run_history` (~94 tokens)

List past and current runs persisted on disk -- survives daemon restarts. Each line is runId, name, status (live / exit N / ?), how long ago it started, and its log size. Pass a runId to read_logs to see its output.

Input parameters:

- `cwd` (string): Project directory; limits the list to that project.
- `limit` (number): Maximum runs to return (default 50).

### `list_network_requests` (~144 tokens)

List HTTP requests the app has made, newest last. One line each: id, method, status (… while in flight), duration in ms, response bytes, URI. Flutter debug/profile sessions only, while the app is running. Captures dart:io HttpClient traffic (package:http and dio with its default adapter) — NOT cupertino_http/cronet_http native clients, WebSockets or raw sockets. Traffic those make is invisible here, not absent.

Input parameters:

- `filter` (string): Case-insensitive regular expression over "METHOD uri".
- `session` (string, required): Session id from list_sessions.
- `tail` (number): How many requests (default 200).

### `get_network_request` (~146 tokens)

One captured request in full: headers, timeline events, and optionally the bodies. Fetched from the running app, so it works only while the session is alive. Flutter debug/profile sessions only, while the app is running. Captures dart:io HttpClient traffic (package:http and dio with its default adapter) — NOT cupertino_http/cronet_http native clients, WebSockets or raw sockets. Traffic those make is invisible here, not absent.

Input parameters:

- `id` (string, required): Request id from list_network_requests (the short number works too).
- `includeBodies` (boolean): Include request and response bodies (capped at 256 KB).
- `session` (string, required)

### `clear_network_requests` (~41 tokens)

Forget every captured request for a session, in the daemon and in the app itself. Useful before reproducing one specific call.

Input parameters:

- `session` (string, required)

### `read_launch_config` (~80 tokens)

Read a project's launch.json exactly as written -- comments, formatting and all -- plus the configurations it defines and any pre-flight problems with them. Returns file: null when the project has none; a file that does not parse still returns its text, with the parse errors.

Input parameters:

- `cwd` (string): Project directory. Defaults to the daemon's current project.

### `write_launch_config` (~148 tokens)

OVERWRITE a project's launch.json with the text given. The whole file is replaced, so read it with read_launch_config first and send the full text back, or comments and configurations you did not mean to touch are lost. Invalid JSONC is refused before anything is written. Returns the path written and any pre-flight issues in the result.

Input parameters:

- `cwd` (string): Project directory. Defaults to the daemon's current project.
- `file` (string): Which convention to create for a project that has neither. A project that already has a launch.json is written back to that same file.
- `text` (string, required): The complete file contents. JSONC: comments and trailing commas are allowed.

### `list_devices` (~39 tokens)

List connected devices, simulators and emulators available to Flutter. Use boot_device to start one that is not running.

Input parameters:

- `cwd` (string)

### `set_debug_flag` (~89 tokens)

Toggle a framework debug flag on a running Flutter session, e.g. ext.flutter.debugPaint with {enabled: true} or ext.flutter.timeDilation with {timeDilation: 5}. Values are sent as the strings Flutter expects.

Input parameters:

- `method` (string, required): Service extension name, e.g. ext.flutter.debugPaint
- `params` (object)
- `session` (string, required)

### `screenshot` (~215 tokens)

Capture the screen of a running session so the visual result of a change can be checked. iOS simulators and Android devices capture the device screen. Web sessions (Vite, Next.js, Flutter web...) load their URL in a headless Chrome, Edge or Brave at the viewport given, and also report the page's console errors and failed requests; they capture at 1x unless scale=2. A capture proves an image was written, not that the UI is right: look at it.

Input parameters:

- `out` (string): Where to save the PNG. Defaults to a path under the daemon's state directory.
- `scale`: Web sessions only: device pixels per CSS pixel. 1 (default) keeps the image small; 2 for fine detail, at about 4x the image tokens.
- `session` (string, required)
- `viewport` (string): Web sessions only: phone (390x844), tablet (820x1180), desktop (1280x800, the default) or WIDTHxHEIGHT.

### `check_change` (~252 tokens)

After editing code: apply the change and return the evidence in one call, instead of hot_reload + wait_for + read_logs + list_network_requests + screenshot. action=auto hot reloads Flutter, leaves web and React Native to HMR / Fast Refresh, and restarts everything else. Answers ok/not ok with only what is new since the call began: reload result, error lines (for web, including the browser console), failed requests, and a screenshot. Prefer this as the default step after every edit.

Input parameters:

- `action` (string): Default auto.
- `scale`: Web sessions only: device pixels per CSS pixel. 1 (default) keeps the image small; 2 for fine detail, at about 4x the image tokens.
- `screenshot` (boolean): Default true. false returns text only; a web page is still loaded headlessly so its console errors count.
- `session` (string, required)
- `settleMs` (number): Wait after the action before collecting evidence (default 1500).
- `viewport` (string): Web sessions only: phone (390x844), tablet (820x1180), desktop (1280x800, the default) or WIDTHxHEIGHT.

### `boot_device` (~97 tokens)

Boot an iOS simulator or Android emulator by name (e.g. "iPhone 17 Pro", "Pixel_8") or id, and wait until it is up. Returns the device id to pass to run_target as deviceId. Call with no name to list what can be booted.

Input parameters:

- `cwd` (string)
- `name` (string): Device name, an unambiguous part of it, or its id.

### `wait_for` (~69 tokens)

Block until a session reaches a state. Use after run_target or hot_restart instead of polling read_logs.

Input parameters:

- `session` (string, required)
- `timeoutMs` (number): Default 60000, capped at 900000 (15 minutes, enough for a first native build).
- `until`

### `session_summary` (~28 tokens)

Cheap structured overview -- call this before deciding what to do next.

Input parameters:

- `session` (string, required)

### `run_proof` (~268 tokens)

Run proof verification and return a zip with screenshots, network call counts, response times and logs.

Input parameters:

- `allow` (string): Comma-separated regexes allowed in noErrors check.
- `appearance` (string): Comma-separated: light,dark
- `branch` (string): Branch to verify in an isolated worktree.
- `checkout` (string): Existing worktree to verify.
- `checks` (string): Comma-separated checks: running,noErrors,noFailedRequests,screenshot
- `cwd` (string)
- `devices` (string): Comma-separated device names, e.g. "iPhone SE,iPhone 16 Pro Max".
- `keep` (boolean): Leave sessions running after the proof.
- `locale` (string): Comma-separated locale codes
- `out` (string): Bundle output directory.
- `route` (string): Deep link or URL to open on each device before capture.
- `settleMs` (number): Milliseconds to wait after navigation before screenshot (default 2000).
- `target` (string, required): Target name or unambiguous substring.
- `textScale` (string): Comma-separated scales, e.g. 1.0,1.5
- `timeoutMs` (number): Per-cell running timeout (default 120000).

### `list_proofs` (~30 tokens)

List past proof bundles, newest first.

Input parameters:

- `limit` (number): Maximum entries (default 20).

### `show_dashboard` (~93 tokens)

Show Baton

Show a live Baton control panel in the chat: every running app with its simulator or URL, status, CPU and memory; per-app Reload, Restart, Pause/Resume, Bring to front, Logs, Screenshot and Stop; quick actions for all apps; and New run (pick a project, target and device, or boot a simulator). Use when the user asks to see or show Baton, their runs or simulators.

### `dashboard_action` (~116 tokens)

Baton card action

Used by the Baton card: refresh; catalog (projects and devices for New run); run a project target on a device; boot a device; logs, reload, restart, pause, resume, front, screenshot, stop or dismiss (forget a finished run) one session; or reload/restart/pause/resume/front/stop all. Returns the card state.

Input parameters:

- `action` (string, required)
- `device` (string)
- `root` (string)
- `session` (string)
- `target` (string)

## Diagnostics

Captured diagnostic sections: Provenance, Dependencies. The full working is on the page: https://verifymcp.io/servers/kanumuri9593-baton/baton-run#diagnostics

## Score history

- 2026-11-05: 91
- 2026-11-04: 91
- 2026-11-03: 90
- 2026-11-02: 90
- 2026-11-01: 89
- 2026-10-31: 89
- 2026-10-30: 88
- 2026-10-29: 88
- 2026-10-28: 88
- 2026-10-27: 87
- 2026-10-26: 87
- 2026-10-25: 86
- 2026-10-24: 86
- 2026-10-23: 85
- 2026-10-22: 85
- 2026-10-21: 81
- 2026-10-20: 81
- 2026-10-19: 81
- 2026-10-18: 81
- 2026-10-17: 81
- 2026-10-16: 81
- 2026-10-15: 81
- 2026-10-14: 81
- 2026-10-13: 81
- 2026-10-12: 50

## Common questions

### What is the Baton MCP server?

Baton is an MCP server listed in the public MCP registry as io.github.kanumuri9593/baton. Run, hot-reload and screenshot Flutter, web, React Native and native iOS/Android apps. This page covers its npm package (baton-run).

### Is the Baton MCP server safe to use?

Baton scores 91 out of 100 on VerifyMCP. We found no known CVEs affecting it as of 5 November 2026. It declares no install or post-install scripts. Its build provenance is signed and verified. That is a record of what we were able to check automatically, not an endorsement. The category breakdown on this page shows every signal behind the number, including the ones we could not confirm.

### What tools does the Baton MCP server expose?

Baton exposes 29 tools: diagnose, run_workflow, inspect_project, list_targets, list_sessions, and 24 more. Their descriptions and schemas cost roughly 2,786 tokens of context every time the server is loaded.

### Is the Baton MCP server still maintained?

Baton is still listed as active in the MCP registry. We last reached this channel on 5 November 2026. Those dates come from our own scans of the registry and the channel itself, not from anything the publisher announced.

### What licence is the Baton MCP server under?

Baton declares the MIT licence, which is OSI-approved. That covers the source only, and says nothing about the cost of any service it calls.

## Links

- npm package: https://www.npmjs.com/package/baton-run
- Socket report: https://socket.dev/npm/package/baton-run
- Repository: https://github.com/kanumuri9593/Baton
- Website: https://kanumuri9593.github.io/Baton/
- Changelog RSS feed: https://verifymcp.io/servers/kanumuri9593-baton/baton-run.xml
- Changelog JSON feed: https://verifymcp.io/servers/kanumuri9593-baton/baton-run.json
- HTML version of this page: https://verifymcp.io/servers/kanumuri9593-baton/baton-run