Spec Driven Development
MCP Server that facilitates spec-driven development workflows, not just Vibe Coding.
Description from the maker (original English)
Not for navigation. Pre-release preview: COLREGS Part C night lights, international text only.
These three are never merged into one score: opinions, measurements and popularity are different things.
We have not written a plain-language summary for this one yet. Below are the facts we read from the registry.
Pick the AI app you use, then copy the settings.
You need
Which app do you use?
Pick your app above. We remember it, so next time the button says “Install in …”.
Open Claude Desktop → Settings → Developer → Edit Config, and paste this into claude_desktop_config.json
{
"mcpServers": {
"colregs": {
"command": "npx",
"args": [
"-y",
"colregs-mcp"
]
}
}
}Run this in your terminal
claude mcp add colregs -- npx -y colregs-mcpPaste into .cursor/mcp.json in your project, or ~/.cursor/mcp.json
{
"mcpServers": {
"colregs": {
"command": "npx",
"args": [
"-y",
"colregs-mcp"
]
}
}
}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.
Paste into .vscode/mcp.json in your project
{
"servers": {
"colregs": {
"type": "stdio",
"command": "npx",
"args": [
"-y",
"colregs-mcp"
]
}
}
}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.
This one runs on your computer, and ChatGPT can only connect to servers on the internet. Pick another app.
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.
Runs on your computer; the tool list appears in your AI app after you install it.
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 8, 2026
The registry name was verified with the GitHub account mark-brannan, which owns the repo
We have not read this repo from GitHub yet
We have not read this repo from GitHub yet
This one runs on your computer; there is no internet endpoint
Runs on your computer and uses local settings instead
Runs on your computer; we can only read tool lists from internet endpoints
Compared with every active server in this category in the registry. Stars count only when the publisher owns the repo.
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.
The registry does not list a required key, though some servers ask you to sign in on first use.
It runs on your computer. Your AI app starts it when needed, and data does not pass through the maker’s servers (unless it calls an online service itself).
The GitHub repo last changed –, and the latest version (0.0.7) reached the registry 3 days ago.
Apps that support MCP, such as Claude Desktop, Claude Code, Cursor and VS Code. You also need Node.js on your computer.
In the same category: Spec Driven Development, Remote Desktop Commander, aamio rendezvous
Servers in the same category from other makers, sorted by stars.
Reading a card:↓ Downloads per weeki★ GitHub starsiThree pillars: reviews · checks · usageiUpdated = latest version published to the registryi
MCP Server that facilitates spec-driven development workflows, not just Vibe Coding.
Remote Desktop Commander: use files, terminal and processes on your own computers from any AI app.
Meet an agent you have not met, exchange messages that expire, and prove it happened. No account.
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.
[!WARNING]
Not for navigation. This is a preview (0.0.x) built on pre-release
rule data. Do not use its output to decide what to do on the water.
The tool surface will change without notice.
An MCP server that answers questions under the COLREGS. Every answer carries its paragraph cites and the verbatim rule text those cites resolve to, so a language model cannot quietly simplify it. The rules are evaluated by colregs-engine; this package only decides how the answer is written down.
Claude Code:
claude mcp add colregs -- npx -y colregs-mcp
Claude connector directory: in Claude, open Settings → Connectors, search for COLREGS and click Connect. No configuration is needed.
Any other MCP client: command npx, arguments -y colregs-mcp, no environment.
From a clone:
git clone https://github.com/mark-brannan/colregs-mcp && cd colregs-mcp && npm install && npm run build
then command node, argument /absolute/path/to/colregs-mcp/dist/cli.js.
| tool | input | returns |
|---|---|---|
evaluate | circumstances | one evaluation: appraisal, signals, indicated, compliance |
rule_text | cite | the verbatim text of a COLREGS paragraph |
light | id | one light's definition: colour, character, arc, placement |
circumstances is the whole picture one vessel is in: self (her own facts), others (each other vessel, with what is known or observed of her) and conditions (visibility, environment). One call evaluates the picture; the response carries four readings of it.
self may show. Each light carries its modality (shall, may, shall-if-practicable) and the paragraph that prescribes it.self on her own, and then against each other vessel: the kind of encounter, whether risk of collision is asserted, and each vessel's role (stand-on, give-way, shall-not-impede).Every id travels with its paragraph cite, and cited_paragraphs holds the verbatim text of every paragraph the response cites. The input vocabulary is generated from the colregs data package at startup. A key or value outside it is an error, never an empty answer.
Takes a paragraph path such as 25(b) or 26(c)(i), a bare rule number such as 25 for every paragraph of that rule, or a range such as 24(a)(ii)-(iv). Returns the International text as transcribed from the USCG amalgamated Navigation Rules, with its source and retrieval date.
Takes a light id as it appears in a response, such as light:sidelights or light:all_round. Returns the Rule 21 definition: colour, character, arc in degrees and placement.
A 12 m sloop under way, alone.
{
"circumstances": { "self": { "fact:propulsion": "propulsion:sail", "fact:activity": "activity:none", "fact:position": "position:underway", "fact:length_m": 11.6 } } }
Trimmed response:
{
"colregs": { "version": "0.3.4" }, "signals": { "applied": [ { "id": "rule:25a", "cite": "25(a)", "modality": "modality:shall" }, { "id": "rule:25b", "cite": "25(b)", "modality": "modality:may" }, { "id": "rule:25c", "cite": "25(c)", "modality": "modality:may" } ], "lawful_displays": { "count": 3, "relation": "any_one_of", "options": [ { "option": 1, "of": 3, "lights": [ { "light": "light:sidelights", "modality": "modality:shall", "prescribed_by": { "id": "rule:25a", "cite": "25(a)" } }, { "light": "light:sternlight", "modality": "modality:shall", "prescribed_by": { "id": "rule:25a", "cite": "25(a)" } } ] }, { "option": 2, "of": 3, "...": "..." }, { "option": 3, "of": 3, "...": "..." } ] }, "modality_key": { "modality:shall": "mandatory", "modality:may": "permitted alternative" }, "cited_paragraphs": { "25(a)": "A sailing vessel underway shall exhibit: (i) sidelights; (ii) a sternlight.", "25(b)": "…", "25(c)": "…" } }, "appraisal": { "self": { "...": "..." }, "others": [] }, "indicated": [] }
The sloop has three lawful displays, because the Rules give her the choice and nothing in the input settles it.
docs/demo.md puts the same question to the same model with and without this server and shows both answers as captured.
Three properties hold for every response:
cited_paragraphs holds the text of every paragraph cited. The only sentences in a response are the Rules' own.shall, may and shall-if-practicable are kept on each light, not summarised per display. modality_key is a one-line glossary of the modalities the response uses.lawful_displays opens with count and relation before any lights appear, and each option numbers itself n of count. There is no top-level lights field to mistake for the answer.A vessel that lawfully shows nothing gets one empty display, not an error. An error is reserved for input outside the vocabulary.
Every response carries colregs.version. The server refuses to start if its own copy of colregs and the engine's are different releases.
colregs-engine and colregs are installed from GitHub main, not from npm releases; the lockfile pins the commit. Both must resolve the same release.
The server collects nothing. It runs as a local stdio process, answers from the colregs data package installed beside it, and makes no network request of its own. There is no telemetry, no logging of what it is asked, and no account. The only network transaction is the install itself, which is npm's.
Automatic from the release tag: npm and the MCP registry. The by-hand steps (Smithery, the Claude directory plugin bundle) are in docs/releasing.md.
AGPL-3.0-only. The fitness-for-navigation disclaimer in colregs carries over.