ci: add CI workflow, fix README badges and remove quest references

.github/workflows/ci.yml

@@ -0,0 +1,53 @@
+name: CI
+
+on:
+  push:
+    branches: [main]
+  pull_request:
+    branches: [main]
+
+jobs:
+  links:
+    name: Check Links
+    runs-on: ubuntu-latest
+
+    steps:
+      - uses: actions/checkout@v4
+
+      - name: Check links in markdown files
+        uses: lycheeverse/lychee-action@v2
+        with:
+          args: >
+            --no-progress
+            --exclude-loopback
+            --accept 200,403,429
+            --timeout 15
+            --retry-wait-time 5
+            --max-retries 3
+            README.md SKILL.md references/*.md
+          fail: true
+        env:
+          GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}
+
+  markdown:
+    name: Markdown Lint
+    runs-on: ubuntu-latest
+
+    steps:
+      - uses: actions/checkout@v4
+
+      - name: Write markdownlint config
+        run: |
+          cat > .markdownlint.json << 'EOF'
+          {
+            "default": true,
+            "MD013": false,
+            "MD033": false,
+            "MD041": false
+          }
+          EOF
+
+      - name: Lint markdown files
+        uses: DavidAnson/markdownlint-cli2-action@v17
+        with:
+          globs: "**/*.md"

README.md

@@ -1,5 +1,8 @@
 # UZPROOF Agent Skill
 
+[![npm](https://img.shields.io/npm/v/%40uzproof%2Fverify?label=npm%3A%20%40uzproof%2Fverify)](https://www.npmjs.com/package/@uzproof/verify)
+[![License: MIT](https://img.shields.io/badge/License-MIT-green.svg)](LICENSE)
+
 Agent Skill for verifying real on-chain usage on Solana. Works with [Claude Code](https://claude.ai/code), [Cursor](https://cursor.com), [GitHub Copilot](https://github.com/features/copilot), [Gemini CLI](https://geminicli.com), and [30+ other AI coding agents](https://agentskills.io/).
 
 ## What is UZPROOF?
@@ -8,6 +11,8 @@ Agent Skill for verifying real on-chain usage on Solana. Works with [Claude Code
 
 UZPROOF is also the first Proof-of-Use attestor on the Solana Attestation Service (SAS), creating permanent on-chain records of verified usage.
 
+**On-chain credential:** [`2chgBfvkwhnHQVVAyXKDK6CBjbCRMQ8aLWrysL5UQyyF`](https://explorer.solana.com/address/2chgBfvkwhnHQVVAyXKDK6CBjbCRMQ8aLWrysL5UQyyF)
+
 ## Install
 
 ```bash

SKILL.md

@@ -14,6 +14,7 @@ You are building with UZPROOF, the first Proof-of-Use verification layer on Sola
 ## Use / Do Not Use
 
 **Use when:**
+
 - Verifying a wallet performed a specific on-chain action (swap, stake, hold, mint)
 - Building reward distribution that requires proof of real usage
 - Gating access to features based on on-chain behavior
@@ -23,6 +24,7 @@ You are building with UZPROOF, the first Proof-of-Use verification layer on Sola
 - Fetching token metadata and live prices
 
 **Do not use when:**
+
 - Executing transactions (use Jupiter, Drift, etc. for that)
 - Reading raw transaction data (use Helius for that)
 - Building wallets or signing transactions
@@ -142,6 +144,7 @@ See [references/attestation.md](references/attestation.md) for SAS integration d
 ## Action Types
 
 ### DeFi Actions
+
 | Type | Verifies |
 |---|---|
 | `defi_swap` | Any swap on a supported DEX |
@@ -164,13 +167,15 @@ See [references/attestation.md](references/attestation.md) for SAS integration d
 | `defi_create_lst` | LST creation |
 
 ### NFT Actions
+
 | Type | Verifies |
 |---|---|
 | `nft_hold` | Owns NFT from collection |
 | `nft_mint` | Minted an NFT |
 | `nft_check` | General NFT ownership check |
 
 ### Utility Actions
+
 | Type | Verifies |
 |---|---|
 | `token_balance` | Raw token balance |
@@ -205,10 +210,10 @@ async function checkEligibility(wallet: string): Promise<boolean> {
 }
 ```
 
-### Quest/Task Verification
+### Task Verification
 
 ```typescript
-async function verifyQuestTask(wallet: string, task: { type: string; config: object }) {
+async function verifyTask(wallet: string, task: { type: string; config: object }) {
   const client = new UzproofClient({ apiKey: process.env.UZPROOF_API_KEY });
   const result = await client.verify({
     wallet,
@@ -231,8 +236,7 @@ const info = await client.detectContract(userProgramId);
 if (info.detected) {
   console.log(`Detected: ${info.program.name} (${info.program.category})`);
   console.log(`Supported verifications: ${info.program.supportedActions.join(', ')}`);
-  // Use suggested quest template
-  const template = info.questTemplate;
+  const template = info.verificationTemplate;
 }
 ```
 
@@ -265,7 +269,7 @@ import { SAS_PROGRAM_ID, UZPROOF_CREDENTIAL, POU_SCHEMA, ACTION_TYPES, SUPPORTED
 
 SAS_PROGRAM_ID     // "22zoJMtdu4tQc2PzL74ZUT7FrwgB1Udec8DdW4yw4BdG"
 UZPROOF_CREDENTIAL // "2chgBfvkwhnHQVVAyXKDK6CBjbCRMQ8aLWrysL5UQyyF"
-POU_SCHEMA         // "8yW2BboQuhp2MMmrQLFz35V6VSqC48MF7wZ5bmzcTeTF"
+POU_SCHEMA         // "9FQiiMtroSHP2Ewqfh3D94GPKnDjmeLT2ftqJ3E7QyWc"
 ACTION_TYPES       // All 24 supported action types
 SUPPORTED_PROTOCOLS // 14 protocol names
 ```

references/anti-fraud.md

@@ -0,0 +1,148 @@
+# Anti-Fraud Scoring
+
+## Overview
+
+Every verification request in UZPROOF runs an anti-fraud check alongside the on-chain data lookup. The check produces a **risk score (0–100)** based on suspicious signals detected on the wallet. It does not block requests — it flags them for review.
+
+The result is returned alongside verification data so your application can decide how to handle high-risk wallets.
+
+## Risk Score
+
+```typescript
+const result = await client.verify({
+  wallet: '7H4RVL...',
+  action: 'defi_swap',
+  config: { program_id: 'JUP6LkbZbjS1jKKwapdHNy74zcZ3tLUZoi5QNyVTaV4' }
+});
+
+// result.riskScore = 0–100
+// result.signals    = array of detected signals
+// result.verified   = true/false (on-chain result, unaffected by risk score)
+```
+
+A score of `0` means no suspicious signals. A score of `100` means maximum risk.
+
+## Signals
+
+### 1. `no_history` — No Transaction History (severity: high, +30)
+
+The wallet has zero recorded transactions on-chain.
+
+```text
+Wallet has zero transaction history
+```
+
+Fresh wallets with no history are a strong indicator of sybil accounts created purely to farm rewards.
+
+### 2. `low_tx_count` — Low Transaction Count (severity: medium, +15)
+
+The wallet has fewer than 5 transactions total.
+
+```text
+Only 2 transactions (min: 5)
+```
+
+Legitimate DeFi users accumulate transactions naturally. Very low counts suggest a wallet created for a single purpose.
+
+### 3. `new_wallet` — Recently Created Wallet (severity: high, +30)
+
+The wallet's oldest transaction is less than 3 days ago.
+
+```text
+Wallet is only 1.2 days old (min: 3)
+```
+
+New wallets are a classic sybil pattern — attackers spin up fresh addresses in bulk before a campaign launches.
+
+### 4. `dust_balance` — Near-Zero SOL Balance (severity: medium, +15)
+
+The wallet holds less than 0.01 SOL.
+
+```text
+SOL balance: 0.0021 SOL (below 0.01 SOL threshold)
+```
+
+Real users need SOL for gas. Dust balances indicate a funded-just-enough-to-transact sybil wallet.
+
+### 5. `fast_completion` — Instant Verification (severity: high, +30)
+
+The wallet was verified less than 10 seconds after the verification flow started.
+
+```text
+Verified 3s after start (min: 10s)
+```
+
+Automated bots complete verifications near-instantly. Humans take longer to read instructions and connect wallets.
+
+## Risk Score Calculation
+
+Scores are additive and capped at 100:
+
+| Severity | Score Added |
+|---|---|
+| `low` | +5 |
+| `medium` | +15 |
+| `high` | +30 |
+| `critical` | +50 |
+
+**Example:** A wallet that is 1 day old (`new_wallet`: +30) and has 2 transactions (`low_tx_count`: +15) gets a risk score of **45**.
+
+## Using Risk Score in Your Application
+
+```typescript
+const client = new UzproofClient({ apiKey: process.env.UZPROOF_API_KEY });
+
+const result = await client.verify({
+  wallet,
+  action: 'defi_swap',
+  config: { program_id: 'JUP6LkbZbjS1jKKwapdHNy74zcZ3tLUZoi5QNyVTaV4' }
+});
+
+if (!result.verified) {
+  // On-chain action not found
+  return { eligible: false, reason: 'action_not_verified' };
+}
+
+if (result.riskScore >= 60) {
+  // High risk — flag for manual review or block
+  return { eligible: false, reason: 'high_fraud_risk', score: result.riskScore };
+}
+
+if (result.riskScore >= 30) {
+  // Medium risk — allow but mark as suspicious
+  return { eligible: true, flagged: true, score: result.riskScore };
+}
+
+// Clean wallet, verified action
+return { eligible: true, flagged: false };
+```
+
+## Recommended Thresholds
+
+| Risk Score | Suggested Action |
+|---|---|
+| 0–29 | Allow |
+| 30–59 | Allow with flag / extra review |
+| 60–100 | Block or require manual review |
+
+Thresholds depend on your campaign's risk tolerance. Airdrops with large rewards should use stricter thresholds than loyalty programs.
+
+## Signal Response Format
+
+```typescript
+// result.signals example
+[
+  {
+    signalType: 'new_wallet',
+    severity: 'high',
+    details: 'Wallet is only 1.2 days old (min: 3)',
+    value: 1.2
+  },
+  {
+    signalType: 'low_tx_count',
+    severity: 'medium',
+    details: 'Only 2 transactions (min: 5)',
+    value: 2
+  }
+]
+```

references/contract-detect.md

@@ -25,7 +25,7 @@ const info = await client.detectContract('JUP6LkbZbjS1jKKwapdHNy74zcZ3tLUZoi5QNy
     "category": "dex",
     "supportedActions": ["defi_swap", "defi_swap_buy", "defi_swap_sell", "defi_swap_volume"]
   },
-  "questTemplate": {
+  "verificationTemplate": {
     "suggestedTitle": "Get started with Jupiter Aggregator v6",
     "suggestedDescription": "Complete tasks to earn XP and prove your on-chain activity with Jupiter Aggregator v6.",
     "suggestedTasks": [
@@ -79,7 +79,7 @@ If a program ID is not in the supported list, `detectContract()` returns:
 {
   "detected": false,
   "program": null,
-  "questTemplate": null
+  "verificationTemplate": null
 }
 ```