Patina

io.github.ramakrishnanhulk20由 GitHub 帳號 ramakrishnanhulk20 發布v1.0.0更新於 Oct 24, 2026

Check how many years of verifiable history a person has behind accounts they already own.

已驗證Streamable HTTP可網頁執行金融安全與監控

概覽

AI 產生的概覽

Patina 提供依據使用者既有帳號歷史計算的 0-100 資歷評分,讓助理能檢查信任門檻並驗證簽章證明。

可以做什麼
依 Patina 使用者名稱查詢完整評分與明細,包括歷史年數與已連結來源。依最低分數、最低年數或兩者同時設定,給出通過或不通過的信任判斷,並附一句可直接引用給使用者的理由。把 GitHub 或 LinkedIn 帳號解析為是否連結公開 Patina 個人頁及其分數,同時不揭露身分。核對評分證明是否確實由 Patina 簽章,比對還原出的地址與公開地址。
適用情境
適合助理需要判斷某個帳號背後有多少可驗證歷史時使用,例如空投、募資輪、投票或試用資格的门檻判斷,或在信任某個帳號前先查核。它衡量的是資歷而非真人唯一性,因此是身分驗證的補充而非替代。
執行需求
遠端 MCP 端點;使用 MCP 工具不需安裝套件、金鑰或 OAuth。README 中的自架方式則需要 Node 20+、Vana 應用程式身分與 VANA_APP_PRIVATE_KEY,以及選用的 Upstash Redis 憑證。
安裝前請注意
評分與證明涉及真實使用者的帳號歷史,查詢時應遵守伺服器聲明的隱私規則。低分不等於指控,老帳號可以被購買,證明在三十天後過期。README 說明 Instagram 已暫停,解析工具不會查詢它。自架需要 Vana 應用程式私鑰並為託管帳戶儲值。

安裝

在 SourceWeft 中

  1. 開啟 儀表板中的 Patina,將其新增到工作區。
  2. 為需要使用其工具的對話啟用該服務。

Web executable,透過 Streamable HTTP。 遠端服務在工作區中設定後即可從網頁執行環境執行。

交給代理程式安裝

把這段提示詞貼上到 Claude Code、Codex、Cursor 等代理程式,或 SourceWeft 對話中。代理程式會讀取該伺服器的安裝指南,先向你展示發布者、端點和登入方式,經你同意後再把它新增到你的客戶端。金鑰不會寫進任何命令。

閱讀 https://sourceweft.com/mcp/io.github.ramakrishnanhulk20%2Fpatina/install.md,並按其中說明把這個 MCP 伺服器接入我的客戶端。修改任何設定之前,先告訴我它的發布者、端點和登入方式,等我確認。不要把 API 金鑰寫進命令或我會提交的檔案;問我金鑰應從哪裡取得。

檢視代理程式所依據的安裝指南

其他 MCP 客戶端

把它新增到你客戶端的 mcpServers 設定中。

{
  "mcpServers": {
    "patina": {
      "type": "http",
      "url": "https://patinadata.xyz/api/mcp"
    }
  }
}

README

[Patina]

Patina

Proof of tenure for accounts a person already owns
Anyone can make a new account. Nobody can make an old one.

Live App • Documentation • MCP Server • Verify a score

Overview • The Score • Privacy • Architecture • Quick Start • Integration • Test Results


Live Deployments

SurfaceURL
Apppatinadata.xyz
Docspatinadata.xyz/docs
MCPpatinadata.xyz/api/mcp
Verifierpatinadata.xyz/verify
Registryio.github.ramakrishnanhulk20/patina
Signing address0x3989bdFaf3BA242d27B4D0cEed98F446d0c52DAD

Overview

Patina reads the history in accounts somebody already owns and turns it into portable, verifiable evidence that a real human has been there for years. It is built on Vana, the data portability protocol: the data lives in the user's own Personal Server, is read once under a grant they can revoke, and never passes through an OAuth handoff that keeps it.

Sybil resistance is unsolved and expensive. An airdrop, a quadratic funding round, a DAO vote or a free trial all need to know whether they are talking to a thousand people or to one person with a thousand wallets. The usual answers are poor: biometrics are invasive, KYC destroys privacy, and most behavioural signals can simply be bought.

Somebody running an account farm can buy followers, bulk-upload posts and fill in a convincing profile. What they cannot buy is a decade.

Why Patina?

KYCWorld ID / proof of personhoodPatina
Proves years of historyNoNoYes
Proves you are one unique personPartlyYesNo
Works without a walletYesNoYes
Works without biometricsYesNoYes
User can revoke access afterwardsNoNoYes
Another app can reuse the proofNoYesYes

Patina answers tenure, not uniqueness. A freshly verified unique human can be produced today; sixteen years of corroborated activity cannot. The two are complementary, and the tool descriptions say so rather than implying otherwise.


Features

Ownership proved, not assumed

Collection is meant to run through Vana Desktop, which opens a browser on the user's own machine and asks them to sign in. Credentials never leave the device and Patina never sees a password.

But Vana has two collection paths and does not report which one it used. The other reads a public page from a profile URL, which proves an account exists and nothing about who holds it. So "desktop only" cannot be asked for; it has to be forced.

Every source therefore requires one scope a public page cannot serve, and the read is refused if it does not come back. LinkedIn must produce connection dates, Spotify saved tracks, YouTube its Watch Later list. A request containing something private cannot be answered from a public profile, so the person ends up on Desktop and has to log in. sources.test.ts fails if a source is ever added without one.

Steam was withdrawn for the same reason. Its connector never signs anybody in: it takes a Steam Web API key and a Steam ID, and a Steam ID is public. Anyone could have handed Patina a stranger's account age and friend dates as their own, and unlike YouTube there was no private scope to require instead.

Instagram is paused for a related reason. Its proof scope, instagram.posts, is one Vana's server collects from the public page, so a read proves the account exists and nothing about who holds it, and Instagram has no private scope Patina could require instead. It is not offered, both Vana routes refuse it before any money is spent, the handle resolver will not look it up, and Instagram data already stored is left out of every score. It comes back when the phone app's signed-in proof lands.

Eight sources, seventeen scopes

GitHub, LinkedIn, Spotify, YouTube, Amazon, Uber, DoorDash and Shop. All of a source's scopes go in one approval, so the whole manifest is eight approval trips rather than seventeen.

Timestamps in, content out

Every scope is requested for its dates. Captions, addresses, emails, track names, game titles and other people's names are discarded before anything is stored, and a test fails if any of them reach the store.

Signed, portable, and dated

Every non-provisional score carries an EIP-191 attestation anyone can verify offline against Patina's published address. No key, no OAuth, CORS open.

Attestations expire after thirty days, and the expiry sits inside the signed message rather than beside it, so checking freshness stays offline too. A revocation list would have fixed the same problem by making every verifier call Patina, which is exactly the dependency people integrate to avoid. Scores are signed with a key separate from the one holding the escrow balance, so a leak of one is not a leak of both.

Answerable about itself

/api/health does a real write-and-read against the database and a live read of the escrow balance, and returns 503 when the deployment cannot serve the next person, so ordinary uptime monitoring catches an outage without being taught what to look for. A password-gated /admin shows the funnel, the user numbers and how many more connections the balance will fund. Funnel counting is done server-side: no third-party script, no cookie banner, nothing that contradicts the privacy page.

Readable by agents

An MCP server exposes the score to AI agents mid-conversation, with a handle resolver that returns tenure without revealing identity.


The Score

Six components, weighted by how expensive each one is in time rather than in money.

ComponentMaxWhat it measures
Age30The oldest date provable across every connected source. Full marks at twelve years.
Continuity25Distinct months the person was actually present for, as absolute count times coverage.
Corroboration15Independent sources agreeing on the date, each weighted by its own age.
Vouches12When other people chose to connect, not how many. Weighted by how long ago.
Depth10Things actually made, discounted when the volume arrived in one burst.
Breadth8Independent accounts backing each other up.

Age and the two time components are earned outright. Depth, Vouches and Breadth are gated behind them: an attacker can manufacture volume, friends and breadth in an afternoon, so those only count to the extent that real elapsed history backs them up. A floor of 15% stops a genuinely young person being flattened to zero for the crime of being nineteen.

Scored against reference profiles, asserted in score.test.ts:

ProfileScore
Eleven active years across four accounts95
Ordinary: eleven years, two accounts, real gaps71
Genuinely young: three years, two accounts, real26
Account farm: 3,900 bought followers, 120 posts in one week2

The gap between an ordinary person and a farm is 69 points. That gap is the product.

Honest limits

Stated on the site itself, not buried here:

  • It cannot prove personhood. It reports how much real history it can see. That is evidence, not a certificate.
  • A low score is not an accusation. Young, private or quiet accounts score low and are doing nothing wrong.
  • Aged accounts can be bought. That turns a free attack into an expensive one whose price climbs with age. It does not make it impossible.
  • Desktop proves session control, not ownership. A bought account arrives with its password. The bar moves from "knows a username" to "holds the credentials", which is a large jump and not the same word.
  • The proof scope is a floor, not a guarantee. Requiring something private forces a real sign-in, and a real sign-in is still only a sign-in. What it rules out is somebody scoring a stranger's public profile, which is the failure that would make the whole number meaningless.

What is read, and what is kept

Every scope below is requested for its timestamps. This table is the trust argument, so it is on the privacy page in roughly this form too.

ScopeKeptDiscarded on arrival
linkedin.connectionsdateConnected[], countnames, headlines, profile URLs
instagram.postsnothing new while Instagram is paused; data already held is not scoredcaptions, images, likes, the entire who_liked[] array
spotify.savedTracksadded_at[], totaltrack names, artists, albums
github.historycreatedAt[], hashed repo, summed engagementPR and issue titles and bodies
linkedin.experience / educationparsed date rangescompanies, job titles, schools, grades
uber.tripsrequestTime[], countpickup and dropoff addresses, fares, cities
amazon / doordash / shop ordersdates, countitems, merchants, totals, delivery addresses
youtube.profilejoinedDate, countsemail address
youtube.watchLaternothing at alleverything; it is requested only as proof of a signed-in session

Timestamps collapse to month buckets before storage. The scorer only ever asks about months, so holding anything finer would be holding it for no reason.


Architecture

Connect flow

mermaid
graph LR    U[User] -->|1 Connect| P[Patina]    P -->|2 createAccessRequest<br/>all scopes for one source| V[Vana Account API]    V -->|3 approval URL| U    U -->|4 Open in Vana Desktop| D[Vana Desktop]    D -->|5 Playwright signs in<br/>on the user's machine| S[Source]    S -->|6 import| PS[Personal Server]    P -->|7 read every scope<br/>one grant| PS    P -->|8 settle 402| E[Escrow Gateway]    P -->|9 normalise, discard, score| R[(Redis)]

Read and settle

mermaid
sequenceDiagram    participant App as Patina    participant PS as Personal Server    participant GW as Escrow Gateway
    App->>PS: GET /v1/data/{scope}    PS-->>App: 402 with accessRecord    App->>GW: authorizeGrantPayment(accessRecord)    GW-->>App: receipt    App->>PS: GET /v1/data/{scope}    PS-->>App: 200 payload    Note over App: repeat per scope, same grant    App->>PS: acknowledgeRead (once, at the end)

acknowledgeRead moves the request to completed, which is terminal and not read-ready. Firing it after the first scope would strand the rest behind a Personal Server no longer serving them, after the user had paid.

Scoring pipeline

mermaid
graph LR    RAW[Raw payload] -->|readScope| F[Fragment<br/>months + counts]    F -->|stored by scope| DB[(Redis)]    DB -->|evidenceFrom| EV[Evidence]    EV -->|scorePatina| SC[Score]    EV -->|buildStory| ST[Story]    SC -->|buildAttestation| AT[Signed attestation]

Fragments are keyed by scope and merged at scoring time. A retried read overwrites its fragment rather than adding its months a second time, so a stuttering connection cannot inflate a score.


Quick Start

Prerequisites

  • Node 20+
  • A Vana app identity from account.vana.org/developers
  • Upstash Redis credentials (optional locally; falls back to an in-memory store)

Installation

bash
git clone https://github.com/ramakrishnanhulk20/Patina.gitcd Patinanpm installcp .env.example .env.local

Fill in VANA_APP_PRIVATE_KEY and VANA_APP_URL, then register the app and fund its escrow:

bash
npm run register:moksha

Development

bash
npm run dev          # development servernpm test             # 118 testsnpm run typecheck    # types, ignoring generated route typesnpm run verify       # typecheck + tests + production build

Integration

REST

Any public score, no key and no OAuth:

bash
curl https://patinadata.xyz/api/verify/ramkumar
json
{  "username": "ramkumar",  "score": 71,  "verdict": "Well established",  "oldestYear": 2012,  "yearsOfHistory": 13.6,  "sourcesConnected": ["github", "linkedin", "spotify"],  "provisional": false,  "issuedAt": "2026-08-26T12:00:00.000Z",  "expiresAt": "2026-09-25T12:00:00.000Z",  "attestation": {    "app": "0x...",    "message": "Patina score attestation
username: ramkumarscore: 71/100...expiresAt: 2026-09-25T12:00:00.000Zapp: 0x...",    "signature": "0x...",    "expiresAt": "2026-09-25T12:00:00.000Z"  }}

Verify an attestation offline

javascript
import { verifyMessage } from "viem";
const ok = await verifyMessage({  address: "0x3989bdFaf3BA242d27B4D0cEed98F446d0c52DAD",  message: attestation.message,  signature: attestation.signature,});

MCP

json
{  "mcpServers": {    "patina": {      "type": "http",      "url": "https://patinadata.xyz/api/mcp"    }  }}
ToolAnswers
get_patina_scoreThe full breakdown for a Patina username
check_thresholdA yes/no trust gate on years, score or source count
resolve_identityA GitHub or LinkedIn handle to a score, without revealing identity. Instagram lookups are paused
verify_attestationWhether a signature is genuinely Patina's

Test Results

174 passing
SuiteTestsCovers
normalize.test.ts30Envelope shapes, discard guarantees, idempotency
score.test.ts27Every component, the gating, the reference profiles
store.test.ts26Cross-device identity, usernames, removing one source, deletion
mcp-lookup.test.ts20Privacy rules on the resolver, signing, verdict drift
story.test.ts19Timeline, activity, exhibits, thin-history copy
routes.test.ts15Real HTTP requests: sessions, the signing floor, the admin gate
device.test.ts12Which devices can finish a connection
attest.test.ts10Signing, expiry, key separation, published-address drift
sources.test.ts9The ownership invariant: no source may be answerable from a public page
ratelimit.test.ts4Escrow abuse limits
altcha.test.ts2Proof-of-work bot check

The one worth knowing about: nothing sensitive survives normalisation feeds in a payload stuffed with a real-looking email address, two home addresses, a caption, a colleague's full name and LinkedIn URL, a game title, an employer, a song, a product and a restaurant, then asserts that none of the thirteen appear anywhere in what gets saved.


Project Structure

src/├── app/│   ├── api/│   │   ├── vana/          request + data: the connect spine│   │   ├── patina/        me, username, data, restore, delete│   │   ├── admin/         session, index rebuild│   │   ├── verify/        public score lookup│   │   ├── badge/         embeddable SVG badge│   │   └── mcp/           MCP server│   ├── connect/           the evidence board, and the phone handoff│   ├── my-data/           see it, download it, remove any of it│   ├── admin/             the funnel, the numbers, the money left│   ├── u/[username]/      public page and story│   ├── verify/            verifier, including an offline checker│   └── docs/              integration reference└── lib/    ├── score.ts           the whole argument, weight by weight    ├── normalize.ts       18 scopes in, timestamps out    ├── sources.ts         the manifest, and the proof scope each source must return    ├── device.ts          whether this device can finish a connection at all    ├── vana.ts            app identity, one controller per source    ├── vana-settle-read.ts  paid reads, escrow settlement    ├── escrow-balance.ts  how many more connections the balance will fund    ├── metrics.ts         the funnel, counted server-side and nowhere else    ├── store.ts           profiles keyed on Personal Server identity    ├── admin.ts           one operator, one password, fails closed    ├── story.ts           evidence as a narrative, and exhibits    ├── attest.ts          signed attestations, expiring, on their own key    └── mcp-lookup.ts      what an agent is allowed to learn

src/lib/score.ts is the whole argument. Every weight carries a comment explaining why it is what it is.


Tech Stack

LayerChoice
FrameworkNext.js 16, React 19, App Router
LanguageTypeScript
StylingTailwind CSS 4, custom token layer
DataVana Data Portability Protocol (@opendatalabs/vana-sdk)
StorageUpstash Redis, in-memory fallback
Signingviem, EIP-191
AgentsModel Context Protocol (mcp-handler)
AbuseALTCHA proof of work, per-session rate limits
Testsnode:test, no framework

License

MIT, see LICENSE.


Acknowledgments

Built on Vana and its Data Portability Protocol. Connectors come from PDP-Connect/data-connectors. Brand glyphs are from Simple Icons and Font Awesome's free brand set, both inlined so the app makes no external request.

來源:README.md,提交 923f721

工具

4

get_patina_score

Get Patina score

Look up a person's full Patina score by their Patina username. Patina reads the age and activity of accounts somebody already owns (YouTube, GitHub, LinkedIn, Spotify and others) through the Vana data portability protocol, and turns that into a 0-100 score plus a signed attestation. Patina scores 0-100 for how much provable history an account holder has, not for how popular or how real-time-verified they are. 80+ (Deeply worn in) means a long, multi-year record corroborated across independent platforms. 60-79 (Well established) is a solid multi-year record. 40-59 (Some real history) is genuine but shorter or thinner. 20-39 (Thin, but genuine so far) is an early account with a little history. Under 20 (Not much to go on yet) looks freshly created, or has almost nothing connected. The inputs, heaviest first: Age (30, full marks at 12 years), Continuity (25, how many separate months the person was actually present for), Corroboration (15, independent sources agreeing on the date), Vouches (12, WHEN other people connected to them rather than how many), Depth (10, things made) and Breadth (8, independent accounts). Follower counts are deliberately worth almost nothing, because they are purchasable. Patina proves TENURE: that whoever holds these accounts has a long, corroborated, verifiable history. It does NOT prove uniqueness or liveness, and it is not a proof-of-personhood check. If you need 'exactly one human, present right now', Patina is the wrong signal and you should say so. Call this when you want the whole picture: the score, the per-component breakdown, how many years of history are provable, which platforms corroborate each other, and an attestation you can verify offline. If you only need a yes or no about whether somebody clears a bar, call check_threshold instead. An unknown username comes back with found: false, which is a normal answer meaning no public Patina profile exists under that name. It is not an error, and it is not evidence against the person.

輸入參數username必填

check_threshold

Check a Patina trust bar

Decide whether somebody clears a trust bar, as a plain true or false. Give a Patina username plus min_score, min_years, or both. At least one is required. Returns `pass`, plus a one-sentence `reason` written to be quoted straight back to a user. This is the right tool for gating decisions: 'has this person been verifiably around for at least five years', 'is this account above 70'. Patina scores 0-100 for how much provable history an account holder has, not for how popular or how real-time-verified they are. 80+ (Deeply worn in) means a long, multi-year record corroborated across independent platforms. 60-79 (Well established) is a solid multi-year record. 40-59 (Some real history) is genuine but shorter or thinner. 20-39 (Thin, but genuine so far) is an early account with a little history. Under 20 (Not much to go on yet) looks freshly created, or has almost nothing connected. The inputs, heaviest first: Age (30, full marks at 12 years), Continuity (25, how many separate months the person was actually present for), Corroboration (15, independent sources agreeing on the date), Vouches (12, WHEN other people connected to them rather than how many), Depth (10, things made) and Breadth (8, independent accounts). Follower counts are deliberately worth almost nothing, because they are purchasable. Patina proves TENURE: that whoever holds these accounts has a long, corroborated, verifiable history. It does NOT prove uniqueness or liveness, and it is not a proof-of-personhood check. If you need 'exactly one human, present right now', Patina is the wrong signal and you should say so. When no profile exists, `pass` is false and `found` is false: treat that as UNKNOWN rather than as a failed check, because most people have never heard of Patina and absence proves nothing about them.

輸入參數username必填min_scoremin_years

verify_attestation

Verify a Patina attestation

Check whether a Patina attestation really was signed by Patina. Every Patina score ships with an EIP-191 signed message; hand this tool that `message` and `signature` and it returns the address it recovered, the address Patina publishes, and whether the two match. IMPORTANT, AND THE WHOLE POINT: you do not need this tool, and where you can, you should not use it. A signature is worth something precisely because ANYONE can check it offline, with no call to Patina, using viem's recoverMessageAddress or ethers' verifyMessage. If you ask Patina's own server whether Patina's own signature is good, you are trusting Patina again and the signature has bought you nothing. That is why this returns the recovered address rather than a bare yes: compare it yourself. Attestations also EXPIRE: the message carries an `expiresAt` line inside the signed bytes, and a genuine signature past that date means the score was true when issued and should be fetched again, not that anybody forged anything.

輸入參數message必填signature必填

resolve_identity

Resolve a platform handle to a Patina score

Find out whether the person behind a platform handle has a Patina score, WITHOUT learning who they are. Supported sources: github, linkedin. Accepts a bare handle, an @handle, or a full profile URL. Returns only three things: whether a public Patina profile is linked to that account, its score, and its years of history. It deliberately never returns the Patina username, and never returns which other platforms that person has connected, because doing so would turn this into a cross-platform de-anonymisation tool. Do not ask it for those; it cannot provide them. Other platforms (youtube, spotify, amazon, uber) are NOT supported here, because for those Patina stores an internal platform id rather than a handle a person could type. Those platforms still count fully toward the score itself; only this lookup is limited. Instagram lookups are paused, because Patina cannot yet prove an Instagram account belongs to the person who connected it. Email addresses are refused outright. Patina scores 0-100 for how much provable history an account holder has, not for how popular or how real-time-verified they are. 80+ (Deeply worn in) means a long, multi-year record corroborated across independent platforms. 60-79 (Well established) is a solid multi-year record. 40-59 (Some real history) is genuine but shorter or thinner. 20-39 (Thin, but genuine so far) is an early account with a little history. Under 20 (Not much to go on yet) looks freshly created, or has almost nothing connected. The inputs, heaviest first: Age (30, full marks at 12 years), Continuity (25, how many separate months the person was actually present for), Corroboration (15, independent sources agreeing on the date), Vouches (12, WHEN other people connected to them rather than how many), Depth (10, things made) and Breadth (8, independent accounts). Follower counts are deliberately worth almost nothing, because they are purchasable.

輸入參數handle必填source必填

版本歷史

1
  1. v1.0.0最新Sep 16, 2026

評分

已滿足適用檢查項的 100%

該評分衡量此收錄條目的完整程度,依據註冊中心條目、端點與公開來源計算得出。它不是對伺服器本身的評價,也不表示伺服器安全或目前可用:可連通性是帶日期的檢查結果,評分也不包含安裝量或使用者評價。

  • 近 30 天內可連通必需25 分
  • 已宣告認證方式必需10 分
  • 已列出工具必需15 分
  • 描述必需15 分
  • 圖示10 分
  • 文件15 分
  • 授權條款10 分