Spec Driven Development
MCP Server that facilitates spec-driven development workflows, not just Vibe Coding.
Description from the maker (original English)
Reusable purchase-requisitions and dated runs of them: pass, fail, not applicable, and a sign-off.
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.
MCP_LICENSE_KEYValues marked secret are keys to your account. Never share them or paste them into a chat.
Which app do you use?
Pick your app above. We remember it, so next time the button says “Install in …”.
Download the maker’s .mcpb file
The link goes to the maker’s own file; we do not host or repackage it. Open the file and Claude Desktop installs it.
Run this in your terminal
Fill in the empty values before running it.
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.
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 theluckystrike, 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.
You need to fill in: MCP_LICENSE_KEY
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 (0.22.8) reached the registry 25 days ago.
Apps that support MCP, such as Claude Desktop, Claude Code, Cursor and VS Code. You also need MCPB on your computer.
In the same category: Spec Driven Development, Remote Desktop Commander, SubA by entendu.
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.
AI-assisted subtitle localization with project rules, glossaries, review and export.
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.
In the [official MCP Registry](https://registry.modelcontextprotocol.io/v0.1/servers/io.github.theluckystrike%2Fpurchase-requisition/versions/latest) (io.github.theluckystrike/purchase-requisition). Checklists you build once and run many times, and the dated record of each run that somebody signs. A purchase-requisition is a named list of steps, optionally grouped into sections, each one required or optional. A run is one pass of that purchase-requisition against a job: every step is marked pass, fail or not applicable, with who marked it and on what day, and a note that says what was found. run_sign_off then puts a name and a date on it and freezes it.
If somebody edits the purchase-requisition afterwards, adds a step or deletes one, every run already in progress keeps the list it started with, and the version it was copied from is recorded on the run.
That is not a caching convenience. A purchase-requisition somebody ticked and signed has to be the list they actually saw. A run that read its steps live from the purchase-requisition would mean a signed handover certificate for eleven checks when the person signing it saw ten, with no field in the record showing that it had happened. It also means deleting a purchase-requisition leaves its runs readable and complete, which is what you want the year afterwards when somebody asks what was checked.
Two smaller rules follow from it:
complete here is a reading: it appears when the last step is answered and goes away again when one is put back to pending.na counts as ANSWERED and never as passed. A step that was looked at and dismissed is a different fact from a step that passed, and merging the two is how a purchase-requisition reports full marks for a job where half the steps did not apply.A required step that is unanswered, a required step that failed, an unanswered optional step, or a run with no steps. force: true signs anyway, and the exceptions stay on the record and print on the report under "Signed with exceptions". They are not lost, and they are not silent.
One-click (.mcpb): download purchase-requisition.mcpb from the latest release and double-click it in Claude Desktop: https://github.com/theluckystrike/mcp-servers/releases/latest
npm publish for @theluckystrike/mcp-purchase-requisition is pending, so the npx line below returns 404 today. Build from source in the meantime; see llms-install.md.
~/Library/Application Support/Claude/claude_desktop_config.json (macOS) or %APPDATA%\Claude\claude_desktop_config.json (Windows):
{
"mcpServers": { "purchase-requisition": { "command": "npx", "args": ["-y", "@theluckystrike/mcp-purchase-requisition"] } } }
claude mcp add purchase-requisition -- npx -y @theluckystrike/mcp-purchase-requisition
~/.cursor/mcp.json (global) or .cursor/mcp.json (project), same entry as Claude Desktop.
| Tool | What it does |
|---|---|
purchase-requisition_create | Create a reusable purchase-requisition: a name, a category, a description |
purchase-requisition_item_add | Add a step: the text, a section heading, and whether it is required |
purchase-requisition_item_remove | Remove a step and bump the version. Runs already under way keep it |
purchase-requisition_show | One purchase-requisition, grouped by section, with a blank printable copy on request |
purchase-requisition_list | Every purchase-requisition with its version, step count and how many runs came from it |
purchase-requisition_delete | Delete a purchase-requisition. Its runs stay readable, because each carries its own copy |
run_start | Start a dated run against a job. The steps are copied into it at this point |
run_check | Mark one step pass, fail or na, with who and when and what was found |
run_show | The run: every step with its answer, the counts, the failures, and what blocks sign-off |
run_list | Runs newest first, filtered by purchase-requisition, status, reference, or only those with failures |
run_sign_off | Sign off with a name and a date, which freezes the run |
run_status | Reopen a complete run, or abandon one when the job did not happen |
run_report | The run as text on every tier. Pro also writes it to out_path as a .txt file |
run_delete | Delete a run. A signed-off one is refused |
license_status | Which tier this install is on and where the key came from |
license_activate | Store a Pro key for this server |
There is also a resource, purchase-requisition://contract, carrying the snapshot rule, the item states, the run status machine, what blocks a sign-off and where this server writes; and a prompt, run_the_purchase-requisition, that walks the whole job in order.
| Free | Pro | |
|---|---|---|
| Checklists you keep | 3 | unlimited |
| Runs of them | unlimited | unlimited |
| Steps per purchase-requisition | up to 500 | up to 500 |
run_show, run_list, the counts | yes | yes |
| The run report as text | yes | yes |
Writing the report to a file with out_path | no | yes |
The meter is on how many DIFFERENT purchase-requisitions you keep, not on how many jobs you check. A trade with one pre-delivery check, one handover sheet and one snag list runs its whole year inside the free tier. Runs are never capped, because capping the running of a purchase-requisition would cap the only thing a purchase-requisition is for. Deleting a purchase-requisition frees a slot.
Get Pro: https://mcp.zovo.one/buy/purchase-requisition (one-time), or all servers for one price at https://mcp.zovo.one/buy/bundle
All data stays local, in ${XDG_DATA_HOME:-~/.local/share}/mcp-servers/purchase-requisition/. There is no network call anywhere in this server, no API key, and no account. The only file it reads that it does not own is the shared business profile, for the name and address at the top of a printed report, and it never writes to it.
Built by theluckystrike. Support: support@zovo.one
Any MCP client (Claude, Cursor, Windsurf, VS Code) can read this repository's documentation directly via GitMCP - no install: