Spec Driven Development
by formulahendry
MCP Server that facilitates spec-driven development workflows, not just Vibe Coding.
Runs on your computer★ 440MITUpdated 14 months ago
Install
Shippo
Overview
Description from the maker (original English)
Multi-carrier shipping for AI agents: compare rates, buy labels, track packages, validate addresses
PopNx MCP’s three pillars
These three are never merged into one score: opinions, measurements and popularity are different things.
- User voiceiNo reviews yetReviews open together with accounts in a later step
- Our basic checksi1 of 1 passedChecked from public information; code not scanned yet
- UsageiNot measured yet★ –
What you can do with it
We have not written a plain-language summary for this one yet. Below are the facts we read from the registry.
- Connection: Runs in the cloud
- Transport: Streamable HTTP
- Category: Other
Install
Pick the AI app you use, then copy the settings.
Before you install
- Basic checks: 1 of 1 passed
- Runs on the maker’s server: what you ask it goes through their server.
- The registry lists no key to fill in.
You need
- Runs in the cloudNothing to install; it connects over the internet
Which app do you use?
Pick your app above. We remember it, so next time the button says “Install in …”.
Works in Claude on the web, desktop and mobile:
- Open Settings → Connectors
- Press Add custom connector
- Paste the address below and press Add
- If it asks you to sign in, sign in with the maker
https://mcp.shippo.comRun this in your terminal
claude mcp add --transport http shippo https://mcp.shippo.comOr copy the settings yourself
Paste into .cursor/mcp.json in your project, or ~/.cursor/mcp.json
{
"mcpServers": {
"shippo": {
"url": "https://mcp.shippo.com"
}
}
}Your browser asks to open the app, then the app shows the settings and asks you to confirm. Key values are left empty for you to fill in.
Or copy the settings yourself
Paste into .vscode/mcp.json in your project
{
"servers": {
"shippo": {
"type": "http",
"url": "https://mcp.shippo.com"
}
}
}Your browser asks to open the app, then the app shows the settings and asks you to confirm. Key values are left empty for you to fill in.
ChatGPT connects only to servers on the internet:
- Open ChatGPT on the web → Settings → Connectors (Apps)
- Press Create or Add (some plans need Developer mode under Advanced first)
- Paste the address below, give it a name and save
- If it asks you to sign in, sign in with the maker
https://mcp.shippo.comMenu names can differ slightly by plan and over time.
Your choice stays on this device in a cookie; no account needed.
These settings are generated from the maker’s registry entry. If the README says otherwise, follow the README.
What it can do
We have not read this server’s tool list yet; we check endpoints in rotation.
Basic checksi
We check information anyone can see; we have not read the code. Checks that do not apply are left out of the count. · Checked Nov 7, 2026
- ✓ Pass
Maker matches the repo
The registry name was verified against the domain shippo.com (the registry checks domain ownership)
- … Not checked yet
Still maintained
We have not read this repo from GitHub yet
- … Not checked yet
Has a licence
We have not read this repo from GitHub yet
- … Not checked yet
Endpoint responds
Not checked yet; we check endpoints in rotation
- … Not checked yet
Sign-in with OAuth
Not checked yet; we check endpoints in rotation
- … Not checked yet
No hidden instructions in tool descriptions
Not checked yet; we read the tool list when we check the endpoint
Compared with others in Otheri
- More stars than 0% of the other 18354 servers in this category
- Published to the registry more recently than 14% of the other 18354 servers in this category
- Basic checks equal to or better than 100% of the other 18354 servers in this category
Compared with every active server in this category in the registry. Stars count only when the publisher owns the repo.
Common questions
Is Shippo safe?
It passed 1 of 1 applicable basic checks. We have not read the code. Before adding a key or token, see what access the server asks for and use a key with only the permissions you need.
What keys or codes does it need?
The registry does not list a required key, though some servers ask you to sign in on first use.
Does it run on my computer or in the cloud?
It runs in the maker’s cloud. Connect over the internet with nothing to install, but what you send passes through the maker’s servers.
Is it still maintained?
The GitHub repo last changed –, and the latest version (1.0.0) reached the registry 4 months ago.
Which apps can use it?
Apps that support MCP, such as Claude Desktop, Claude Code, Cursor and VS Code. It connects by URL, so apps that accept remote MCP servers work too, such as ChatGPT in developer mode.
What can I use instead?
In the same category: Spec Driven Development, Remote Desktop Commander, Qemu MCP
Alternatives
Servers in the same category from other makers, sorted by stars.
Reading a card:★ GitHub starsiThree pillars: reviews · checks · usageiUpdated = latest version published to the registryi
Remote Desktop Commander
by desktop-commander
Remote Desktop Commander: use files, terminal and processes on your own computers from any AI app.
Runs in the cloud★ 211Updated 2 days ago
Install
Qemu MCP
by 0xmortuex
Boot and drive QEMU VMs from an AI agent: keys, screenshots, serial. No guest agent.
Runs on your computerUpdated today
Install
README from the maker
Open the README
Copied from the maker’s repo as written. PopNx MCP has not reviewed or endorsed it. HTML, images and scripts are removed. Links lead off this site.
Shippo AI
This repo is the one-stop shop for building AI-powered shipping integrations with Shippo.
It contains:
- 9 Agent Skills: Workflow knowledge for AI assistants covering rate shopping, address validation, label purchase (with customs), package tracking, batch shipping, shipping cost analysis, support-ticket drafting, integration best practices, and SDK/API upgrades. Authored once and distributed across multiple AI surfaces.
- Claude Code Plugin (
providers/claude/plugin/): Install via--plugin-diror the plugin marketplace (/plugin marketplace add goshippo/ai). - OpenAI Codex Plugin (
providers/codex/plugin/): Install via the Codex plugin marketplace; bundles the skills plus the OAuth MCP server. - ClawHub Skill (
providers/clawhub/skills/shippo/): Install viaopenclaw skills install @shippo/shippo. - Claude apps (claude.ai / Desktop / Cowork): The whole plugin is packaged as a single upload-ready ZIP (
shippo-plugin.zip), attached to every GitHub Release. One upload provisions all the skills. - Knowledge Pack (ChatGPT and other chat assistants) (
providers/knowledge-pack/shippo-knowledge-pack.md): A single consolidated markdown for assistants that do not loadSKILL.mdfolders. A user drops it into a chat, a Custom GPT's Knowledge, or a Project as context. It provides the shipping knowledge; live actions still use the hosted MCP connector.
What is a skill?
A skill is a folder containing a SKILL.md file, YAML frontmatter (at minimum: name and description) plus markdown instructions that tell an AI assistant how to perform a specific task. Skills can also bundle reference docs, scripts, and templates.
rate-shopping/├── SKILL.md # required: metadata + instructions └── README.md # optional: human-facing orientation
Agents load skills by progressive disclosure in three stages:
- Discovery: at startup, the agent loads only each skill's
nameanddescription, just enough to know when it might be relevant. - Activation: when a user prompt matches a skill's description, the agent loads the full
SKILL.mdbody into context. - Execution: the agent follows the instructions, optionally loading referenced files (
shippo/references/*.md) as it works.
Agent Skills is an open standard originally developed by Anthropic. The same SKILL.md works in Claude Code, Cursor, OpenAI Codex, GitHub Copilot, VS Code, and 30+ other agents.
In this repo, the 9 skills under skills/ are the canonical source. They're propagated into providers/claude/plugin/skills/ and providers/codex/plugin/skills/ (1:1 mirrors), providers/clawhub/skills/shippo/ (consolidated digest), and providers/knowledge-pack/shippo-knowledge-pack.md (a single upload-ready knowledge pack for ChatGPT and other non-skill-loading assistants) automatically by the sync scripts.
Model Context Protocol (MCP)
Shippo hosts a remote MCP server with per-user OAuth. Each user authorizes once through Shippo, there is no API key to copy. The Claude Code and OpenAI Codex plugins point at this endpoint and trigger the sign-in on first use.
| URL | Transport | Auth |
|---|---|---|
https://mcp.shippo.com | Streamable HTTP | Per-user Shippo OAuth |
For per-tool semantics and usage, see the Shippo MCP server docs.
Building on OpenAI? See Using the Shippo MCP from the OpenAI Responses API / Agents SDK for the developer config (no submission required).
Capabilities
The 9 skills in this repo are organized by mode of engagement: what the user is doing, not by product surface. The AI assistant matches the user's intent to one of three modes, then loads the right skill.
Decide, "where do I start?"
| Skill | What it does |
|---|---|
shippo-best-practices | Decision-router for Shippo integrations, which API to use, test vs. live mode discipline, response handling, critical rules |
Do, "execute this workflow"
| Skill | What it does |
|---|---|
address-validation | Validate, parse, and standardize US and international addresses |
rate-shopping | Compare rates across USPS, UPS, FedEx, DHL, and 30+ carriers |
label-purchase | Purchase domestic and international shipping labels with customs handling |
tracking | Track packages across carriers with status history, substatus codes, and webhooks |
batch-shipping | Process CSV files of shipments and generate labels in bulk |
shipping-analysis | Analyze costs, optimize package dimensions, compare carriers, review historical spend |
shippo-support-ticket | Build an auto-classified, routing-tagged support ticket (human + JSON) for a single shipment or label; read-only, for Shippo support agents |
Maintain, "upgrade or migrate"
| Skill | What it does |
|---|---|
upgrade-shippo | Guide for upgrading SDK versions, MCP server updates, breaking-change migration |
A user who already knows the workflow they need ("buy a label", "track this package") jumps straight to a Do skill. A user starting fresh ("I'm building a checkout flow with shipping, where do I start?") hits the Decide skill, which routes them to the right Do skill. Maintenance gets its own skill so production-readiness questions don't compete with workflow content.
The 9 skills lean on 11 shared reference docs under skills/shippo/references/ (carriers, customs, CSV format, error reference, etc.). Skills load references on demand, the AI doesn't pull all 11 into context, just the ones a given workflow needs.
Install
Claude Code
git clone https://github.com/goshippo/ai.gitclaude --plugin-dir ./ai/providers/claude/plugin
Or install from the plugin marketplace:
/plugin marketplace add goshippo/ai/plugin install shippo@shippo
On first use, run /mcp, select the Shippo server, and sign in to authorize the MCP over OAuth (no API key to copy).
Skills are namespaced under /shippo:: invoke directly with /shippo:rate-shopping, /shippo:label-purchase, /shippo:tracking, etc., or just describe what you're doing in natural language.
OpenAI Codex
Codex installs the Shippo plugin (skills + OAuth MCP) from this repo's plugin marketplace:
codex plugin marketplace add goshippo/aicodex plugin add shippo@shippo # install the "shippo" plugin codex mcp login shippo # authorize the remote MCP over OAuth
See providers/codex/plugin/ for details. (To pull just the skill content without the plugin, Codex's skill-installer can also install a single providers/codex/plugin/skills/ directory.)
ClawHub
openclaw skills install @shippo/shippo(Published as @shippo/shippo on the ClawHub registry.)
Claude apps (claude.ai / Desktop / Cowork)
The Claude apps load the plugin as a single ZIP. shippo-plugin.zip (the whole plugin: manifest, OAuth MCP config, and all skills) is attached to every GitHub Release. Download it and add it via the app's Plugins UI. A Team/Enterprise admin can provision it org-wide in one step: Organization settings → Plugins → upload shippo-plugin.zip → set "Installed by default" (or assign to a group), and all skills become available to members. (Code execution must be enabled in Organization settings.)
To build the ZIP locally: npm run build:app-plugin (output in dist/app-plugin/).
Shippo account
You'll need a Shippo account. Getting rates and validating addresses incur no charge; purchasing a label uses Shippo's discounted carrier rates and charges your account. The Claude Code and Codex plugins authorize per-user via OAuth on first use, so there's no API key to copy.
How it works
This plugin bundles two things, with a deliberate division of labor between them:
- Skills (this repo): Cross-tool workflow narrative: routing decisions (checkout vs single label vs batch), UX gates ("ask before purchasing a live-mode label"), CSV ingestion, validation sequencing, test/live mode discipline, response-handling rules. Loaded on activation when the user's request matches a skill's description.
- MCP server (docs): Per-tool semantics: tool name, parameters, return shape, single-call constraints. Each tool description is terse, one verb phrase, one tool. Workflow guidance is intentionally NOT duplicated here.
The skills teach the assistant *how* to ship across multiple API calls. The MCP server gives the assistant the *per-call* truth about each tool. The two surfaces are disjoint by design, same precedent Stripe uses (terse mcp.stripe.com tool descriptions, rich stripe/agents skills): so raw MCP users get accurate per-tool semantics and skill-installed users additionally get the workflow narrative, without contradiction.
Repo layout
skills/: canonical skill content (9 skills + 11 shared references). Edit here; everything else flows from here.providers/claude/plugin/: Claude Code plugin distribution. 1:1 mirror of canonical viascripts/sync.js.providers/codex/plugin/: OpenAI Codex plugin.skills/is a 1:1 mirror of canonical viascripts/sync.js;.codex-plugin/plugin.json+.mcp.json(hand-authored) carry the manifest and the OAuth MCP wiring. Cataloged from.agents/plugins/marketplace.jsonat the repo root.providers/clawhub/skills/shippo/: ClawHub bundle distribution. TheSKILL.mdis auto-generated fromSKILL.md.template(hand-curated framing) + canonical skill bodies viascripts/compose-clawhub-digest.js. References are auto-synced viascripts/build-clawhub-bundle.js.dist/app-plugin/: the singleshippo-plugin.zipfor the Claude apps, built fromproviders/claude/plugin/byscripts/build-app-plugin.js(not committed; produced on demand and on release).scripts/: sync, compose, and build helpers.
Authoring
# 1. Edit canonical contentvim skills/ /SKILL.md
(or skills/shippo/references/ .md, or providers/clawhub/skills/shippo/SKILL.md.template
if you're changing ClawHub-only framing)
2. Sync + verify (one command)
npm test
3. Commit canonical edits AND synced output together
git add -A && git commit -m "..."
npm test runs all the sync steps (Claude Code + Codex mirrors, ClawHub digest compose, ClawHub references sync) and verifies the result is internally consistent. CI runs the same command. No npm install needed, the repo has no third-party dependencies, just scripts.
Preview your edit
- Claude Code: run
claude --plugin-dir ./providers/claude/pluginfrom the repo root to launch Claude Code with the local plugin loaded. Edits toskills/ /SKILL.mdare reflected immediately. Skills are namespaced under/shippo:(e.g.,/shippo:rate-shopping). - ClawHub digest: after
npm testruns, the rendered output lives atproviders/clawhub/skills/shippo/SKILL.md: read it directly to see what ClawHub-installed users will get. There's no local-server preview today.
See CONTRIBUTING.md for the full authoring discipline, including the version-bump rules and the cross-reference phrasing rule for skill content.
License
MIT
Only the beginning is shown; read the rest in the repo