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
Agiscorecard Web3 Protocol
Overview
Description from the maker (original English)
Read dated x402, ERC-8004 and token-interface profiles with official sources and freshness limits.
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://protocol.agiscorecard.com/mcpRun this in your terminal
claude mcp add --transport http agiscorecard-web3-protocol https://protocol.agiscorecard.com/mcpOr copy the settings yourself
Paste into .cursor/mcp.json in your project, or ~/.cursor/mcp.json
{
"mcpServers": {
"agiscorecard-web3-protocol": {
"url": "https://protocol.agiscorecard.com/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.
Or copy the settings yourself
Paste into .vscode/mcp.json in your project
{
"servers": {
"agiscorecard-web3-protocol": {
"type": "http",
"url": "https://protocol.agiscorecard.com/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.
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://protocol.agiscorecard.com/mcpMenu 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 with the GitHub account f-tiger, which owns the repo
- … 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 32686 servers in this category
- Published to the registry more recently than 59% of the other 32686 servers in this category
- Basic checks equal to or better than 100% of the other 32686 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 Agiscorecard Web3 Protocol 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 (1.7.0) reached the registry 27 days 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, Microsoft Exchange Online 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
Microsoft Exchange Online MCP
by servosity
Exchange Online admin for MSPs: forwarding audits, EWS posture and offboarding, no PowerShell
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.
Web3 Workbench
Ten independent, usable free-beta tools and one hub. Source and generated fixtures are shared; every host serves only its own pages. Existing Agent Delivery Lab remains a separate deployment.
Hub: https://web3.agiscorecard.com/
| Host | Tool | Implemented task |
|---|---|---|
| reconcile.agiscorecard.com | Stable Reconcile | Declared invoice/transfer exceptions |
| evidence.agiscorecard.com | Agent Evidence | Task/version evidence, deduplication, uncertainty |
| route.agiscorecard.com | Route Lab | Bounded fallback sequence planning |
| protocol.agiscorecard.com | Protocol Ledger | Five dated profiles, free API/OpenAPI |
| permit.agiscorecard.com | Data Permit | Declared grant compatibility and proposed splits |
| compute.agiscorecard.com | Compute Lens | Cost per accepted workload output |
| incentives.agiscorecard.com | Incentive Lab | Organic economics and capped rewards |
| proof.agiscorecard.com | Proof Plan | zkML cost, latency and artifact manifest |
| calls.agiscorecard.com | Call Lens | Strict decoding of supported token calls |
| disclosures.agiscorecard.com | RWA Notes | Dated disclosure comparisons |
Development and release
Node.js 22+ and Python 3 with Pillow and DejaVu Sans fonts are required. Browser/offline calculation engines have no npm dependencies; the remote MCP server uses the pinned official SDK.
python3 -m pip install -r requirements-build.txtnpm ci npm test python3 scripts/check_html.py npx --yes wrangler@4 deploy --dry-run
npm test builds 55 HTML pages, 30 worked scenarios, per-host metadata and share images, and the offline bundle, then runs 71 tests. WEB3_BUILD_REVISION pins the actual checked-out commit. Generated dist/, release module and credential-free generated Wrangler config are not committed.
The path-scoped GitHub workflow validates, checks all hostname ownership, binds the existing database, deploys the Worker, and checks every production host and asset against the build. A collision stops before publication. Existing credentials remain in GitHub Secrets. Branch runs validate only; main may deploy. Do not substitute the trigger SHA for the checkout SHA after the stale-main guard.
Only web3_studio_feedback is created/read/updated. No other database table is modified. Optional feedback is constrained enums; IP is used only as the transient rate-limit key. QA samples are explicitly flagged and excluded from demand. Grouped submission counts never represent verified individuals. TTL cleanup occurs on feedback/stats requests and may be delayed during inactivity.
Product limits
Browser inputs are local. No model provider, wallet, live-chain simulation, transfer, subscription or proof verification is connected. Documentation states each tool's own limits. Protocol records are editorial summaries, not complete schemas or compatibility certification. New algorithms mean reproducible combinations of known methods here, not a novel cryptographic or economic theorem.
Research and falsification conditions: portfolio research. Every tool's catalog includes alternatives, scope, commercial hypothesis and a concrete validation condition. Public downloads are limited to the explicit offline bundle and its license; no repository archive is served.
Experience, discovery and sharing (1.1)
Each tool has three runnable scenarios, per-record-group CSV paste, same-tab result comparison, Markdown reports and print/PDF output. Public share links allow only known scenario IDs; private inputs never appear in generated URLs. Clearing/reloading the tab removes in-memory records and comparisons.
Search titles describe actual tasks. Human-readable methods and examples back the structured WebApplication, FAQ and breadcrumb data. No fabricated ratings, usage statistics or rich-result promises. guide.md, tool.json, input.schema.json, llms.txt and llms-full.txt are navigation/integration aids, not special eligibility requirements for AI search. The schema is structural; the engine enforces its additional constraints.
/publish.html on the hub provides ten share cards and optional branded links. Distributed HTML card links use rel="nofollow". No automatic outreach, paid links or third-party endorsements are claimed.
The existing weekly/manual IndexNow workflow validates the key and the canonical page set on each of the 11 hosts before submitting. It never submits on push. node scripts/indexnow.mjs --check validates without submitting. A 200/202 submission response is not evidence of indexing or ranking.
Release smoke verification uses asset-manifest.generated.json to hash-check the exact assets for each host; generated files are excluded from git. See growth implementation record.
MCP (1.2)
Unified endpoint: https://web3.agiscorecard.com/mcp. Each tool host also exposes /mcp, scoped to its own calculator. Streamable HTTP, no key; official SDK 2.0.0 handles 2026-07-28 per-request negotiation and stateless 2025-era compatibility. GET/DELETE return 405; persistent subscriptions are not offered. The existing AGI Scorecard main-domain MCP remains separate.
The hub exposes ten deterministic calculators plus search and fetch. All results preserve the existing algorithms, exact amounts, release revision, methodology citation, date and limits. resources/list exposes public method references, capabilities and fictional inputs. No arbitrary resource fetching or private report publication is implemented. Annotations declare read-only, non-destructive, idempotent, closed-world tools.
Remote MCP arguments are sent to the server, unlike browser/offline inputs. Application code neither persists them nor logs their contents or sends them to a model/source URL. Hosting and the connecting client's own policies still apply. No D1 writes occur in MCP. A separate 120/minute/IP binding limits calls across the 11 hosts; body size is at most 128 KiB including the protocol envelope. Missing rate-limit binding fails closed. Exact Origin validation rejects unknown browser origins; server-side clients without Origin remain supported.
/for-agents.html is the human setup/citation/privacy guide. /.well-known/mcp.json is explicitly a descriptive metadata file, not a standard auto-install mechanism or official registry listing. MCP availability is not evidence of AI search citations.
Tests use the official 2.0 client pinned to 2026-07-28 and the official 1.30 client for 2025-era compatibility. scripts/mcp-smoke.mjs repeats negotiation, tool discovery, resources, all 30 scenarios, citations and host scoping against the deployed endpoint. CI installs the lockfile, retains the stale-main guard and uses existing Cloudflare credentials. No new scheduled workflow is introduced.
Market context and conversion measurement (1.3)
The hub now serves source-stamped /market.html, /briefs.html, three task guides, /api/market, /api/briefs and /updates.xml. Price snapshots refresh on request after five minutes; official metadata after one hour. Stale/unavailable data is explicit and data older than a day is hidden. These are venue references, not executable quotes or verified payments. Public source polling does not change any browser or MCP calculation contract.
Optional, off-by-default fixed-category usage measurement writes only the isolated web3_studio_events table. /api/growth excludes QA unless ?qa=1; reports count opt-in events, not people, linked sessions, verified backlinks or revenue. A separate MEASURE_LIMIT binding caps submissions. /api/stats feedback is separate. Run node scripts/market-smoke.mjs after deploying and node scripts/growth-audit.mjs --check for source/discovery/measurement health. The existing daily fleet-autopilot retains 30 daily diagnostic records.
Research board (1.4)
/market.html now combines eight curated assets, daily history comparisons, source-linked research questions and a local evidence notebook. /api/research is a separate read-only feed; the original four-asset /api/market contract and deterministic MCP tools remain available. status describes snapshot retrieval; every quote has its own freshness status. Changes use daily samples within two hours of the intended interval, never synthetic candles. Missing history does not erase usable current prices.
Watchlist baselines and notes are written to browser storage only on explicit Save, expire after 30 days and have separate Delete actions. They are never uploaded. Export is a user-authored Markdown file, not automatic publication or an AI verdict. Four optional fixed-category events support evaluating the experiment without collecting notes, saved prices or asset selections. scripts/research-smoke.mjs verifies actual source/history connectivity after deployment; the existing daily audit covers this feed too.
Read-only aggregate coverage (2026-10-11)
GET and HEAD /api/growth and /api/stats execute SELECT statements only. Schema setup and 35-day retention cleanup happen on accepted event/feedback submissions, never during reporting. Missing storage or tables return HTTP 503 with status: "unavailable" and groups, daily, today set to null. An empty array is only returned after a successful query. Internal errors and raw inputs are not included in public responses.
Both endpoints include generatedAt, asOf, timezone: "UTC", scope, window, and metric definitions. groups retains its legacy dimensions: 28 rolling days for growth; 35 rolling days of feedback for stats. daily contains sparse date groups for the previous 28 complete UTC days; today is separate and partial. Every window states inclusive start and exclusive end. Every row includes site; the hub queries all sites without duplicating totals onto each tool host. A tool hostname returns only its own rows. Growth's qa=1 is QA-only and uncached, while default growth excludes QA. Feedback retains the qa dimension in all groups: business reports must filter qa=0 explicitly. Self-reported own_completed does not establish an independently verified task. Counts are events/submissions, not users, MCP demand, registrations or revenue.
Successful public reads advertise Cache-Control: public, max-age=3600; unavailable responses and growth QA reads use no-store. generatedAt records query generation, not a claim that source data is fully processed. This change adds no new event collection, D1 tables, persistent visitor IDs or scheduled job. For a deployment that must avoid synthetic business activity, set FLEET_READ_ONLY_PROBE=1: smoke scripts verify assets, metadata and sources but skip feedback/event submissions, aggregate reads, and MCP tools/call requests. The reporting collector can then fetch each aggregate exactly once.