mess-mcp

Name: mess-mcp
Author: diegosouzapw

AI Community

Repo stars 39
Author updated Live
Author repo awesome-omni-skill

Fluxly profile Facts only: domain, agents, trust score, runtime, permissions and network

Domain: AI
Compatible agents: Claude Code
Cursor
Cline
Codex
Windsurf
Gemini CLI

+20
Trust score: 88 / 100 · community maintained
Author / version / license: @diegosouzapw · no license declared
Token usage: Heavy
Setup complexity: Manual integration
External API key: Required · GitHub
Operating systems: macOS · Linux · Windows
Runtime requirements: Node.js · Python
Permissions: Read-only
Write / modify
Shell exec
Network behavior: External requests
Install commands: 26 variants

Profile is derived at build time from SKILL.md and install vectors. Subject to drift from author intent.

Heads up: 未限定 allowed-tools，默认拥有全部工具权限。

Output preview mess-mcp.preview

---
name: mess-mcp
description: MCP server tools for creating and managing MESS physical-world task requests. Provides mess, mes…
category: ai
runtime: Node.js / Python
---

# mess-mcp output preview

## PART A: Task fit
- Use case: MCP server tools for creating and managing MESS physical-world task requests. Provides mess, mess_status, mess_capabilities, mess_request, mess_answer, mess_cancel, and mess_fetch tools..
- Inputs: target material, constraints, expected output, and acceptance criteria.
- Evidence boundary: follow “Overview / Tools / mess - Send MESS Protocol Messages” and do not present inference as author intent.

## PART B: Execution result
- **01** The card summarizes the use case; runtime output centers on “MCP server tools for creating and managing MESS physical-world task requests. Provides mess, mess_status, mess_capabilities, mess_request, mess_answer, mess_cancel, and mess_fetch tools.”.
- **02** When the source has headings, the agent prioritizes “Overview / Tools / mess - Send MESS Protocol Messages” so the result follows the author’s structure.
- **03** Typical output includes task judgment, concrete steps, required commands or file edits, validation, and follow-up options.
- **04** Risk context follows the fingerprint: read files, write/modify files, run shell commands; may access external network resources; requires GitHub API keys.

## Running Rules
- read files, write/modify files, run shell commands; may access external network resources; requires GitHub API keys.
- Validate with a small sample before expanding scope.
- Return the result, validation criteria, and next iteration options.

SKILL.md SKILL.md

---
name: mess-mcp
description: MCP server tools for creating and managing MESS physical-world task requests. Provides mess, mes…
category: ai
source: diegosouzapw/awesome-omni-skill
---

# mess-mcp

## When to use
- MCP server tools for creating and managing MESS physical-world task requests. Provides mess, mess_status, mess_capabil…
- Use it when the task has clear inputs, repeatable steps, and validation criteria.

## What to provide
- Target material, scope, expected result, and forbidden changes.
- Whether network, commands, file writes, or external services are allowed.

## Execution rules
- Organize steps around “Overview / Tools / mess - Send MESS Protocol Messages” and keep inference separate from source facts.
- read files, write/modify files, run shell commands; may access external network resources; requires GitHub API keys.
- Validate with a small sample before expanding the task.

## Output requirements
- Return the deliverable, key evidence, validation method, and next action.
- Mark missing information as unknown; do not invent commands, platforms, or dependencies.

Evidence boundary and execution chain

The author source anchors workflow facts; repository files anchor sources and commands; Fluxly only adds fit, limitations, and quality judgment.

skill "mess-mcp" {
  input      -> user goal + target files + boundaries + acceptance criteria
  context    -> Overview / Tools / mess - Send MESS Protocol Messages
  rules      -> SKILL.md triggers / order / output contract
  runtime    -> Node.js / Python | read files, write/modify files, run shell commands | may access external network resources
  guardrails -> requires GitHub API keys + small-sample validation + diff/log review
  output     -> copyable result + checklist + next iteration
}

Interpretation is structured for decision-making; original keeps the upstream SKILL.md unchanged.

Decide Fit First

Core job: MCP server tools for creating and managing MESS physical-world task requests. Provides mess, mess_status, mess_capabilities, mes…
Best fit: Use it when the task has reusable inputs, steps, and validation criteria rather than a one-off answer.
Avoid forcing it: If the source lacks commands, platform support, or external-service evidence, keep those fields unknown instead of guessing.

Design Intent

Structure: The skill is organized around “Overview”, “Tools”, “mess - Send MESS Protocol Messages”, “messstatus - Check Request Status”, showing how the author expects the agent to judge fit, collect context, and produce verifiable output.
Trigger evidence: Prioritize the author’s wording around when to use it, what context to collect, and what output shape to produce.
Evidence boundary: Author text states facts, repository files prove commands and paths, and Fluxly only adds fit, limits, and usage judgment.

How To Use It

Inputs: Provide target material, scope, expected result, forbidden changes, and validation method.
Invocation: Name mess-mcp directly; if the source includes slash commands, start with the command and then add task context.
Validation: Start small and check whether the result follows “Overview / Tools / mess - Send MESS Protocol Messages” before expanding.

Boundaries And Review

Dependencies: Prepare GitHub API keys before running a full task.
Permissions: Declared permissions include read / write / shell-exec; ask the agent to state file, command, and rollback boundaries before acting.
Quality bar: A useful result names the deliverable, evidence, and next action. Generic prose means the task needs tighter context.

MESS MCP Server

Overview

The MESS MCP Server provides Claude Desktop with tools to create and manage physical-world task requests. It syncs with GitHub (or local files) to store request threads.

Tools

`mess` - Send MESS Protocol Messages

Create new requests or update existing ones using free-form YAML.

Input: YAML-formatted MESS message

Why YAML? The mess tool accepts raw MESS protocol messages, allowing you to:

Include any valid MESS protocol fields, not just common ones
Add custom context, metadata, or instructions
Construct complex multi-part messages
Use the full expressiveness of the MESS protocol

The examples below show common patterns, but you can include any fields defined in the MESS protocol specification.

Creating a Request

- v: 1.0.0
- request:
    intent: Check if the garage door is closed
    context:
      - Getting ready for bed
      - Want to make sure house is secure
    priority: normal
    response_hint:
      - image
      - text

Response:

ref: "2026-02-01-001"
status: pending
message: "Request created: 2026-02-01-001"

Updating a Request

Send a status update:

- status:
    re: "2026-02-01-001"
    code: claimed

Cancel a request:

- cancel:
    re: "2026-02-01-001"
    reason: "No longer needed"

`mess_status` - Check Request Status

Query pending requests or get details on a specific thread.

List All Active Requests

ref: null

Response:

- ref: "2026-02-01-001"
  status: pending
  intent: Check if the garage door is closed
  requestor: claude-desktop
  executor: null
  priority: normal
  created: "2026-02-01T22:00:00Z"
  hasUpdates: true    # changed since last mess_status call

- ref: "2026-02-01-002"
  status: claimed
  intent: What's in the fridge?
  requestor: claude-desktop
  executor: teague-phone
  priority: normal
  hasUpdates: false   # no changes since last check

Note: The hasUpdates flag indicates whether the thread changed since your last mess_status call. New threads always have hasUpdates: true. Use mess_wait instead of polling if you want to wait for changes efficiently.

Get Specific Thread

ref: "2026-02-01-001"

Response (v2.1.0 format):

ref: "2026-02-01-001-garage-check"
status: completed
intent: Check if the garage door is closed
requestor: claude-desktop
executor: teague-phone
messages:
  - from: claude-desktop
    received: "2026-02-01T22:00:00Z"
    MESS:
      - v: "1.0.0"
      - request:
          id: garage-check
          intent: Check if the garage door is closed

  - from: teague-phone
    received: "2026-02-01T22:05:00Z"
    re: "2026-02-01-001-garage-check"      # message-level reference
    MESS:
      - status:
          code: completed
      - response:
          content:
            - image:
                resource: "content://2026-02-01-001-garage-check/att-002-image-door.jpg"
                mime: "image/jpeg"
                size: 245891
            - "All clear - garage door is closed and locked"

attachments:
  - name: att-002-image-door.jpg
    resource: "content://2026-02-01-001-garage-check/att-002-image-door.jpg"

Note: Images are returned as content:// resource URIs instead of inline base64 to keep responses lightweight. Use mess_fetch to fetch attachment content when needed.

Configuration

GitHub Mode (Recommended)

{
  "mcpServers": {
    "mess": {
      "command": "node",
      "args": ["/path/to/mcp/index.js"],
      "env": {
        "MESS_GITHUB_REPO": "username/mess-exchange",
        "MESS_GITHUB_TOKEN": "github_pat_xxxxx",
        "MESS_GITHUB_ONLY": "true",
        "MESS_AGENT_ID": "claude-desktop"
      }
    }
  }
}

Local Mode

{
  "mcpServers": {
    "mess": {
      "command": "node",
      "args": ["/path/to/mcp/index.js"],
      "env": {
        "MESS_DIR": "~/.mess",
        "MESS_AGENT_ID": "claude-desktop"
      }
    }
  }
}

Hybrid Mode (Local + GitHub Sync)

{
  "mcpServers": {
    "mess": {
      "command": "node",
      "args": ["/path/to/mcp/index.js"],
      "env": {
        "MESS_DIR": "~/.mess",
        "MESS_GITHUB_REPO": "username/mess-exchange",
        "MESS_GITHUB_TOKEN": "github_pat_xxxxx",
        "MESS_AGENT_ID": "claude-desktop"
      }
    }
  }
}

`mess_capabilities` - Discover Available Capabilities

List physical-world capabilities that executors in this exchange can perform.

Input: Optional tag filter

List All Capabilities

# No input needed

Response:

- id: camera
  description: Take and attach photos
  tags: [attachments]
- id: check-door
  description: Check if doors are locked or closed
  tags: [security, physical-access]
- id: hands
  description: Has human hands for physical manipulation
  tags: [physical-access]

Filter by Tag

tag: security

Response:

- id: check-door
  description: Check if doors are locked or closed
  tags: [security, physical-access]
- id: check-stove
  description: Verify stove/oven is turned off
  tags: [security, safety]

Use this to understand what kinds of tasks you can request. Capabilities are defined in capabilities/*.yaml in the exchange.

`mess_request` - Create Request (Simple)

A simpler alternative to the raw mess tool for creating requests.

Input:

intent: "Check if the garage door is closed"
context:
  - "Getting ready for bed"
priority: elevated
response_hints:
  - image

Response:

ref: "2026-02-01-001"
status: pending
message: "Request created"

`mess_answer` - Answer Question

Respond to an executor's question when status is needs_input.

Input:

ref: "2026-02-01-001"
answer: "The living room ceiling light, not the lamp"

`mess_cancel` - Cancel Request

Cancel a pending or in-progress request.

Input:

ref: "2026-02-01-001"
reason: "No longer needed"  # optional

`mess_wait` - Wait for Changes

Wait for changes to threads instead of polling mess_status repeatedly.

Input:

ref: "2026-02-01-001"  # optional: wait for specific thread
timeout: 60             # optional: seconds (default: 60, max: 43200 / 12 hours)

Response:

updated:
  - ref: "2026-02-01-001"
    status: claimed
    hasUpdates: true
waited: 5
timedOut: false

Adaptive polling: Poll frequency scales with timeout (~30 polls total):

60s timeout → polls every 2s
5 min → every 10s
1 hour → every 2 min
12 hours → every 24 min

When to use:

After creating a request, call mess_wait to be notified when claimed/completed
Returns immediately if thread changed since wait started
Use for efficient long-running waits instead of polling

Example workflow:

# 1. Create a request
mess_request:
  intent: "Check if the garage door is closed"

# 2. Wait for it to be completed (up to 1 hour)
mess_wait:
  ref: "2026-02-01-001"
  timeout: 3600

# 3. Response returns when status changes or timeout

`mess_fetch` - Fetch Resource Content

This is how you retrieve images and attachments. When mess_status returns content:// URIs, use this tool to fetch the actual content.

Input:

uri: "content://2026-02-01-001/photo.jpg"

Supported URI schemes:

content://{ref}/{filename} - Attachments (images, files)
thread://{ref} - Full thread data (envelope + messages)
thread://{ref}/envelope - Thread metadata only
thread://{ref}/latest - Most recent message
mess://help - This documentation

Response for images:

uri: "content://2026-02-01-001/photo.jpg"
mimeType: "image/jpeg"
encoding: "base64"
data: "/9j/4AAQSkZJRg..."

Response for thread data: Returns the thread content as YAML.

Example workflow:

# 1. Check status, see there's an image
mess_status:
  ref: "2026-02-01-001"
# Response includes: resource: "content://2026-02-01-001/photo.jpg"

# 2. Fetch the image
mess_fetch:
  uri: "content://2026-02-01-001/photo.jpg"
# Response includes base64 image data

Request Fields

Field	Required	Description
`intent`	Yes	What you need done (be specific)
`id`	No	Your local identifier for tracking (exchange assigns canonical `ref`)
`context`	No	List of relevant context strings
`priority`	No	`background`, `normal`, `elevated`, `urgent`
`response_hint`	No	Expected response types: `text`, `image`, `video`, `audio`

Free-Form Extensions

The MESS protocol is extensible. Beyond the standard fields above, you can include:

- v: 1.0.0
- request:
    intent: Check the garden irrigation system
    context:
      - Haven't watered in 3 days
    priority: normal
    # Standard fields above, custom fields below:
    location: backyard
    equipment_needed:
      - hose access
      - manual valve knowledge
    safety_notes:
      - Watch for wasps near the shed
    deadline: before 6pm today

Custom fields are preserved in the thread and visible to executors. Use them for domain-specific context that doesn't fit the standard fields.

Status Codes

Status	Meaning
`pending`	Waiting for executor to claim
`claimed`	Executor is working on it
`in_progress`	Executor actively working
`needs_input`	Executor needs clarification
`completed`	Task finished successfully
`failed`	Could not complete
`declined`	Executor declined the request
`cancelled`	Request was cancelled

Common Patterns

Create and Wait

# 1. Create request
mess:
  - v: 1.0.0
  - request:
      intent: Is anyone home?

# 2. Periodically check status
mess_status:
  ref: "2026-02-01-001"

# 3. When completed, read the response

Urgent Request

- v: 1.0.0
- request:
    intent: Did I leave the stove on?
    context:
      - Just left the house
      - Can't remember if I turned it off
    priority: urgent
    response_hint:
      - image
      - text

Follow-up After needs_input

# Original request got needs_input status
# Executor asked: "Which light?"

# Send clarification (v2.1.0: answer block, message-level re: handled automatically)
- answer:
    id: light-clarification
    value: "The living room ceiling light, not the lamp"

Resources

The MCP server uses content:// and thread:// URIs for attachments and thread data.

Fetching Attachments

When mess_status returns a thread, images and files are referenced as resource URIs:

image:
  resource: "content://2026-02-01-001/att-002-image-door.jpg"
  mime: "image/jpeg"
  size: 245891

To fetch the actual image, use mess_fetch:

mess_fetch:
  uri: "content://2026-02-01-001/att-002-image-door.jpg"

This keeps status responses lightweight and avoids blowing up context with large base64 payloads.

Resource URI Format

content://{thread-ref}/{attachment-filename}

Examples:

content://2026-02-01-001-garage-check/att-002-image-door.jpg
content://2026-02-01-003-fridge/att-005-image-contents.jpg

Thread Resources

Read thread data with attachments automatically rewritten to content:// URIs:

thread://{ref}          # Full thread (envelope + messages)
thread://{ref}/envelope # Just metadata (status, executor, history)
thread://{ref}/latest   # Most recent message only

Example - Read full thread:

# Request: thread://2026-02-01-001

# Response:
envelope:
  ref: "2026-02-01-001"
  status: completed
  intent: Check the garage door
messages:
  - from: claude-desktop
    MESS:
      - request:
          intent: Check the garage door
  - from: teague-phone
    MESS:
      - response:
          content:
            - image:
                resource: "content://2026-02-01-001/att-001-door.jpg"
            - "Door is closed"

Example - Read just envelope:

# Request: thread://2026-02-01-001/envelope

# Response:
ref: "2026-02-01-001"
status: completed
executor: teague-phone
updated: "2026-02-01T22:05:00Z"

Error Handling

Thread not found:

error: "Thread 2026-02-01-999 not found"

Missing required field:

error: "Missing re: field"

GitHub API error:

error: "GitHub API error: 401"

File Locations

macOS

Config: ~/Library/Application Support/Claude/claude_desktop_config.json
Local MESS dir: ~/.mess/

Linux

Config: ~/.config/claude/claude_desktop_config.json
Local MESS dir: ~/.mess/

Windows

Config: %APPDATA%\Claude\claude_desktop_config.json
Local MESS dir: %USERPROFILE%\.mess\

Install command

One-click installPick a runner, click Copy, paste into your terminal, press enter

npx skills add diegosouzapw/awesome-omni-skill/skills/tools/mess-mcp -y

Open general install guide

View skill source

Works with

Claude Code
Cursor
Cline
Codex
Windsurf
Gemini
OpenCode
Roo Code
AAide
+Augment
GitHub Copilot
OpenClaw
KKiro
Junie
WWarp
Goose
Trae
Kilo Code
OpenHands
Amp
Zencoder
CContinue
AIAider
CodeBuddy
CodeGeeX
Replit Agent

Security audit

88 / 100

Community

未声明 license/许可。
未声明 version。
未限定 allowed-tools，默认拥有全部工具权限。

Metadata

Source: github
diegosouzapw/awesome-omni-skill
中文

More by @diegosouzapw

Category & tags

Category AI

Decide Fit First

Design Intent

How To Use It

Boundaries And Review

MESS MCP Server

Overview

Tools

mess - Send MESS Protocol Messages

Creating a Request

Updating a Request

mess_status - Check Request Status

List All Active Requests

Get Specific Thread

Configuration

GitHub Mode (Recommended)

Local Mode

Hybrid Mode (Local + GitHub Sync)

mess_capabilities - Discover Available Capabilities

List All Capabilities

Filter by Tag

mess_request - Create Request (Simple)

mess_answer - Answer Question

mess_cancel - Cancel Request

mess_wait - Wait for Changes

mess_fetch - Fetch Resource Content

Request Fields

Free-Form Extensions

Status Codes

Common Patterns

Create and Wait

Urgent Request

Follow-up After needs_input

Resources

Fetching Attachments

Resource URI Format

Thread Resources

Error Handling

File Locations

macOS

Linux

Windows

Discussion

`mess` - Send MESS Protocol Messages

`mess_status` - Check Request Status

`mess_capabilities` - Discover Available Capabilities

`mess_request` - Create Request (Simple)

`mess_answer` - Answer Question

`mess_cancel` - Cancel Request

`mess_wait` - Wait for Changes

`mess_fetch` - Fetch Resource Content