thumbgate 1.0.0 → 1.1.0

package/.claude-plugin/marketplace.json

@@ -1,6 +1,6 @@
 {
   "name": "thumbgate",
-  "version": "1.0.0",
+  "version": "1.1.0",
   "plugins": [
     {
       "name": "thumbgate",

package/.claude-plugin/plugin.json

@@ -1,7 +1,7 @@
 {
   "name": "thumbgate",
   "description": "Pre-action gates that block AI coding agents from repeating known mistakes. Captures feedback, auto-promotes failures into prevention rules, and enforces them via PreToolUse hooks.",
-  "version": "1.0.0",
+  "version": "1.1.0",
   "author": {
     "name": "Igor Ganapolsky"
   },

package/.well-known/mcp/server-card.json

@@ -1,6 +1,6 @@
 {
   "name": "thumbgate",
-  "version": "1.0.0",
+  "version": "1.1.0",
   "description": "ThumbGate — 👍👎 feedback that teaches your AI agent. Thumbs down a mistake, it never happens again.",
   "homepage": "https://github.com/IgorGanapolsky/thumbgate",
   "transport": "stdio",

package/adapters/README.md

@@ -3,7 +3,7 @@
 - `chatgpt/openapi.yaml`: import into GPT Actions.
 - `gemini/function-declarations.json`: Gemini function-calling definitions.
 - `mcp/server-stdio.js`: underlying local MCP stdio server implementation.
-- `claude/.mcp.json`: example Claude Code MCP config using `npx --yes --package thumbgate@1.0.0 thumbgate serve`.
+- `claude/.mcp.json`: example Claude Code MCP config using `npx --yes --package thumbgate@1.1.0 thumbgate serve`.
 - `codex/config.toml`: example Codex MCP profile section using the same version-pinned portable launcher.
 - `amp/skills/thumbgate-feedback/SKILL.md`: Amp skill template.
 - `opencode/opencode.json`: portable OpenCode MCP profile using the same version-pinned portable launcher.

package/adapters/claude/.mcp.json

@@ -2,13 +2,13 @@
   "mcpServers": {
     "thumbgate": {
       "command": "npx",
-      "args": ["--yes", "--package", "thumbgate@1.0.0", "thumbgate", "serve"]
+      "args": ["--yes", "--package", "thumbgate@1.1.0", "thumbgate", "serve"]
     }
   },
   "hooks": {
     "preToolUse": {
       "command": "npx",
-      "args": ["--yes", "--package", "thumbgate@1.0.0", "thumbgate", "gate-check"]
+      "args": ["--yes", "--package", "thumbgate@1.1.0", "thumbgate", "gate-check"]
     }
   }
 }

package/adapters/codex/config.toml

@@ -1,9 +1,9 @@
 # Codex MCP profile (copy into ~/.codex/config.toml or merge section)
 [mcp_servers.thumbgate]
 command = "npx"
-args = ["--yes", "--package", "thumbgate@1.0.0", "thumbgate", "serve"]
+args = ["--yes", "--package", "thumbgate@1.1.0", "thumbgate", "serve"]
 
 # Hard PreToolUse hook for Codex
 [hooks.pre_tool_use]
 command = "npx"
-args = ["--yes", "--package", "thumbgate@1.0.0", "thumbgate", "gate-check"]
+args = ["--yes", "--package", "thumbgate@1.1.0", "thumbgate", "gate-check"]

package/adapters/mcp/server-stdio.js

@@ -97,6 +97,7 @@ const {
   assembleUnifiedContext,
   formatUnifiedContext,
 } = require('../../scripts/context-manager');
+const { exportHfDataset } = require('../../scripts/export-hf-dataset');
 
 const PRO_CHECKOUT_URL = 'https://thumbgate-production.up.railway.app/checkout/pro';
 
@@ -118,7 +119,7 @@ const {
   finalizeSession: finalizeFeedbackSession,
 } = require('../../scripts/feedback-session');
 
-const SERVER_INFO = { name: 'thumbgate-mcp', version: '1.0.0' };
+const SERVER_INFO = { name: 'thumbgate-mcp', version: '1.1.0' };
 const COMMERCE_CATEGORIES = [
   'product_recommendation',
   'brand_compliance',
@@ -493,6 +494,14 @@ async function callToolInner(name, args) {
     case 'export_dpo_pairs':
       enforceLimit('export_dpo');
       return buildExportDpoResponse(args);
+    case 'export_hf_dataset': {
+      enforceLimit('export_dpo');
+      const outputDir = args.outputDir ? resolveSafePath(args.outputDir) : undefined;
+      return toTextResult(exportHfDataset({
+        outputDir,
+        includeProvenance: args.includeProvenance !== false,
+      }));
+    }
     case 'export_databricks_bundle': {
       enforceLimit('export_databricks');
       const outputPath = args.outputPath ? resolveSafePath(args.outputPath) : undefined;

package/adapters/opencode/opencode.json

@@ -7,7 +7,7 @@
         "npx",
         "--yes",
         "--package",
-        "thumbgate@1.0.0",
+        "thumbgate@1.1.0",
         "thumbgate",
         "serve"
       ],

package/config/mcp-allowlists.json

@@ -22,6 +22,7 @@
       "bootstrap_internal_agent",
       "prevention_rules",
       "export_dpo_pairs",
+      "export_hf_dataset",
       "export_databricks_bundle",
       "construct_context_pack",
       "evaluate_context_pack",

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "thumbgate",
-  "version": "1.0.0",
+  "version": "1.1.0",
   "description": "ThumbGate — Make your AI coding agent self-improving. Every mistake becomes a prevention rule that physically blocks the agent from repeating it. Feedback-driven enforcement via PreToolUse hooks, Thompson Sampling for adaptive gates, SQLite+FTS5 lesson DB, and LanceDB vector search. Your agent gets smarter with every session.",
   "homepage": "https://thumbgate-production.up.railway.app",
   "repository": {
@@ -66,7 +66,7 @@
     "social:post-everywhere:dry": "node scripts/post-everywhere.js --dry-run",
     "social:reply-monitor": "node scripts/social-reply-monitor.js",
     "social:reply-monitor:dry": "node scripts/social-reply-monitor.js --dry-run",
-    "test": "npm run test:schema && npm run test:loop && npm run test:dpo && npm run test:kto && npm run test:api && npm run test:proof && npm run test:e2e && npm run test:rlaif && npm run test:attribution && npm run test:quality && npm run test:intelligence && npm run test:training-export && npm run test:deployment && npm run test:operational-integrity && npm run test:workflow && npm run test:billing && npm run test:cli && npm run test:watcher && npm run test:autoresearch && npm run test:ops && npm run test:tessl && npm run test:gates && npm run test:evoskill && npm run test:gates-hardening && npm run test:workers && npm run test:social-analytics && npm run test:memalign && npm run test:xmemory-lite && npm run test:filesystem-search && npm run test:zernio && npm run test:obsidian-export && npm run test:lesson-db && npm run test:lesson-rotation && npm run test:memory-dedup && npm run test:feedback-quality && npm run test:sync-version && npm run test:check-congruence && npm run test:tool-registry && npm run test:feedback-to-rules && npm run test:memory-firewall && npm run test:belief-update && npm run test:hosted-config && npm run test:cloudflare-sandbox && npm run test:mcp-config && npm run test:plan-gate && npm run test:pulse && npm run test:semantic-layer && npm run test:data-pipeline && npm run test:optimize-context && npm run test:principle-extractor && npm run test:analytics-window && npm run test:funnel-analytics && npm run test:experiment-tracker && npm run test:build-metadata && npm run test:context-engine && npm run test:hf-papers && npm run test:marketing-experiment && npm run test:seo-gsd && npm run test:verify-run && npm run test:export-dpo-pairs && npm run test:license && npm run test:bot-detector && npm run test:postinstall && npm run test:funnel-invariants && npm run test:cli-telemetry && npm run test:pro-parity && npm run test:model-tier-router && npm run test:computer-use-firewall && npm run test:skill-exporter && npm run test:statusline && npm run test:evolution && npm run test:org-dashboard && npm run test:multi-hop-recall && npm run test:synthetic-dpo && npm run test:thumbgate-skill && npm run test:learn-hub && npm run test:feedback-fallback && npm run test:metaclaw && npm run test:server-lock && npm run test:control-tower && npm run test:pii-scanner && npm run test:data-governance && npm run test:lesson-inference && npm run test:lesson-retrieval && npm run test:reflector-agent && npm run test:feedback-session && npm run test:feedback-history-distiller && npm run test:hallucination-detector && npm run test:history-distiller && npm run test:predictive-insights && npm run test:prove-predictive-insights && npm run test:statusbar-cli && npm run test:generate-instagram-card && npm run test:instagram-thumbgate-post && npm run test:publish-instagram-thumbgate && npm run test:lesson-synthesis && npm run test:background-governance && npm run test:memory-migration && npm run test:prompt-dlp && npm run test:ephemeral-store && npm run test:agent-security && npm run test:skill-progressive && npm run test:per-step-scoring && npm run test:weekly-auto-post && npm run test:social-quality-gate && npm run test:a2ui-engine && npm run test:gate-satisfy && npm run test:money-watcher && npm run test:utm && npm run test:product-feedback && npm run test:feedback-root-consolidator && npm run test:engagement-audit && npm run test:install-growth-automation && npm run test:publish-thumbgate-launch && npm run test:reconcile-thumbgate-campaign && npm run test:reddit-publisher && npm run test:schedule-thumbgate-campaign && npm run test:social-reply-monitor && npm run test:sync-launch-assets",
+    "test": "npm run test:schema && npm run test:loop && npm run test:dpo && npm run test:kto && npm run test:api && npm run test:proof && npm run test:e2e && npm run test:rlaif && npm run test:attribution && npm run test:quality && npm run test:intelligence && npm run test:training-export && npm run test:deployment && npm run test:operational-integrity && npm run test:workflow && npm run test:billing && npm run test:cli && npm run test:watcher && npm run test:autoresearch && npm run test:ops && npm run test:tessl && npm run test:gates && npm run test:evoskill && npm run test:gates-hardening && npm run test:workers && npm run test:social-analytics && npm run test:memalign && npm run test:xmemory-lite && npm run test:filesystem-search && npm run test:zernio && npm run test:obsidian-export && npm run test:lesson-db && npm run test:lesson-rotation && npm run test:memory-dedup && npm run test:feedback-quality && npm run test:sync-version && npm run test:check-congruence && npm run test:tool-registry && npm run test:feedback-to-rules && npm run test:memory-firewall && npm run test:belief-update && npm run test:hosted-config && npm run test:cloudflare-sandbox && npm run test:mcp-config && npm run test:plan-gate && npm run test:pulse && npm run test:semantic-layer && npm run test:data-pipeline && npm run test:optimize-context && npm run test:principle-extractor && npm run test:analytics-window && npm run test:funnel-analytics && npm run test:experiment-tracker && npm run test:build-metadata && npm run test:context-engine && npm run test:hf-papers && npm run test:marketing-experiment && npm run test:seo-gsd && npm run test:verify-run && npm run test:export-dpo-pairs && npm run test:export-hf-dataset && npm run test:license && npm run test:bot-detector && npm run test:postinstall && npm run test:funnel-invariants && npm run test:cli-telemetry && npm run test:pro-parity && npm run test:model-tier-router && npm run test:computer-use-firewall && npm run test:skill-exporter && npm run test:statusline && npm run test:evolution && npm run test:org-dashboard && npm run test:multi-hop-recall && npm run test:synthetic-dpo && npm run test:thumbgate-skill && npm run test:learn-hub && npm run test:feedback-fallback && npm run test:metaclaw && npm run test:server-lock && npm run test:control-tower && npm run test:pii-scanner && npm run test:data-governance && npm run test:lesson-inference && npm run test:lesson-retrieval && npm run test:reflector-agent && npm run test:feedback-session && npm run test:feedback-history-distiller && npm run test:hallucination-detector && npm run test:history-distiller && npm run test:predictive-insights && npm run test:prove-predictive-insights && npm run test:statusbar-cli && npm run test:generate-instagram-card && npm run test:instagram-thumbgate-post && npm run test:publish-instagram-thumbgate && npm run test:lesson-synthesis && npm run test:background-governance && npm run test:memory-migration && npm run test:prompt-dlp && npm run test:ephemeral-store && npm run test:agent-security && npm run test:skill-progressive && npm run test:per-step-scoring && npm run test:weekly-auto-post && npm run test:social-quality-gate && npm run test:a2ui-engine && npm run test:gate-satisfy && npm run test:money-watcher && npm run test:utm && npm run test:product-feedback && npm run test:feedback-root-consolidator && npm run test:engagement-audit && npm run test:install-growth-automation && npm run test:publish-thumbgate-launch && npm run test:reconcile-thumbgate-campaign && npm run test:reddit-publisher && npm run test:schedule-thumbgate-campaign && npm run test:social-reply-monitor && npm run test:sync-launch-assets",
     "test:feedback-fallback": "node --test tests/feedback-fallback.test.js",
     "test:metaclaw": "node --test tests/metaclaw-features.test.js",
     "test:server-lock": "node --test tests/server-stdio-lock.test.js",
@@ -105,6 +105,8 @@
     "test:seo-gsd": "node --test tests/seo-gsd.test.js",
     "test:verify-run": "node --test tests/verify-run.test.js",
     "test:export-dpo-pairs": "node --test tests/export-dpo-pairs.test.js",
+    "test:export-hf-dataset": "node --test tests/export-hf-dataset.test.js",
+    "export:hf": "node scripts/export-hf-dataset.js",
     "seo:gsd": "node scripts/seo-gsd.js plan",
     "seo:gsd:write": "node scripts/seo-gsd.js plan --write",
     "test:congruence": "node scripts/check-congruence.js",

package/plugins/claude-codex-bridge/.claude-plugin/plugin.json

@@ -1,6 +1,6 @@
 {
   "name": "codex-bridge",
-  "version": "1.0.0",
+  "version": "1.1.0",
   "description": "Run Codex review, adversarial review, and second-pass handoffs from Claude Code while keeping ThumbGate reliability memory in the loop.",
   "author": {
     "name": "Igor Ganapolsky",

package/plugins/claude-codex-bridge/.mcp.json

@@ -5,7 +5,7 @@
       "args": [
         "--yes",
         "--package",
-        "thumbgate@1.0.0",
+        "thumbgate@1.1.0",
         "thumbgate",
         "serve"
       ]

package/plugins/codex-profile/.codex-plugin/plugin.json

@@ -1,6 +1,6 @@
 {
   "name": "codex-profile",
-  "version": "1.0.0",
+  "version": "1.1.0",
   "description": "ThumbGate for Codex: pre-action gates, skill packs, hallucination detection, PII scanning, progressive disclosure (82% token savings), and MCP-backed reliability memory.",
   "author": {
     "name": "Igor Ganapolsky",

package/plugins/codex-profile/.mcp.json

@@ -5,7 +5,7 @@
       "args": [
         "--yes",
         "--package",
-        "thumbgate@1.0.0",
+        "thumbgate@1.1.0",
         "thumbgate",
         "serve"
       ]

package/plugins/codex-profile/INSTALL.md

@@ -31,7 +31,7 @@ The following block is appended to `~/.codex/config.toml`:
 ```toml
 [mcp_servers.thumbgate]
 command = "npx"
-args = ["--yes", "--package", "thumbgate@1.0.0", "thumbgate", "serve"]
+args = ["--yes", "--package", "thumbgate@1.1.0", "thumbgate", "serve"]
 ```
 
 The repo-local Codex app plugin ships the same runtime path through `plugins/codex-profile/.mcp.json`, so the manual config and plugin metadata stay aligned.

package/plugins/codex-profile/README.md

@@ -29,7 +29,7 @@ That profile launches:
 ```toml
 [mcp_servers.thumbgate]
 command = "npx"
-args = ["--yes", "--package", "thumbgate@1.0.0", "thumbgate", "serve"]
+args = ["--yes", "--package", "thumbgate@1.1.0", "thumbgate", "serve"]
 ```
 
 ## Why this exists

package/plugins/cursor-marketplace/.cursor-plugin/plugin.json

@@ -2,7 +2,7 @@
   "name": "thumbgate",
   "displayName": "ThumbGate",
   "description": "👍👎 Thumbs down a mistake — your AI agent won't repeat it. Thumbs up good work — it remembers the pattern.",
-  "version": "1.0.0",
+  "version": "1.1.0",
   "author": {
     "name": "Igor Ganapolsky"
   },

package/plugins/opencode-profile/INSTALL.md

@@ -25,7 +25,7 @@ The portable profile adds this MCP server entry:
   "mcp": {
     "thumbgate": {
       "type": "local",
-      "command": ["npx", "--yes", "--package", "thumbgate@1.0.0", "thumbgate", "serve"],
+      "command": ["npx", "--yes", "--package", "thumbgate@1.1.0", "thumbgate", "serve"],
       "enabled": true
     }
   }

package/public/index.html

@@ -66,7 +66,9 @@ __GA_BOOTSTRAP__
     "Background Agent Governance — per-agent pass rates, CI auto-feedback",
     "Memory Migration — imports Claude Code MEMORY.md into unlimited SQLite DB",
     "Prompt-Level DLP — scans tool call inputs before execution",
-    "Per-Step Scoring — every gate decision becomes a DPO/KTO training signal"
+    "Per-Step Scoring — every gate decision becomes a DPO/KTO training signal",
+    "HuggingFace Export — share PII-redacted agent traces as open training datasets",
+    "Unified Context — one-call context assembly with session, lessons, guards, and code-graph"
   ],
   "offers": [
     {
@@ -578,7 +580,7 @@ __GA_BOOTSTRAP__
 <!-- HOW IT WORKS -->
 <section class="how-it-works" id="how-it-works">
   <div class="container">
-    <div class="section-label">New in v1.0.0</div>
+    <div class="section-label">New in v1.1.0</div>
     <h2 class="section-title">Three steps to stop repeated AI failures</h2>
     <div class="steps">
       <div class="step">
@@ -670,6 +672,7 @@ __GA_BOOTSTRAP__
           <li>All MCP integrations (Claude Code, Cursor, Codex, etc.)</li>
           <li>PreToolUse hook blocking</li>
           <li>Local SQLite lesson DB</li>
+          <li>Unified context assembly — one call gets session, lessons, guards, and code-graph</li>
           <li><a href="/guide" style="color:var(--cyan);text-decoration:underline;">Setup guide for all agents →</a></li>
         </ul>
         <a href="https://www.npmjs.com/package/thumbgate" target="_blank" rel="noopener" class="btn-free">Install Free</a>
@@ -698,6 +701,7 @@ __GA_BOOTSTRAP__
           <li><a href="/dashboard" style="color:var(--cyan);text-decoration:underline;">Visual gate debugger →</a> see every blocked action and the gate that fired so you can trust the system in minutes</li>
           <li>Auto-connect — activate once with your license key, then your running agents appear automatically on your local dashboard</li>
           <li><a href="/dashboard" style="color:var(--cyan);text-decoration:underline;">DPO training data export →</a> turn real thumbs-downs into ready-to-use preference pairs for fine-tuning (LoRA / JSONL)</li>
+          <li><strong>HuggingFace dataset export</strong> — share PII-redacted agent traces as open training datasets (<code>npm run export:hf</code>)</li>
           <li><strong>Model Hardening Advisor</strong> — get recommendations on when and how to fine-tune your model to natively avoid recurring failures</li>
           <li>Personal local dashboard — every Pro user gets a localhost dashboard without extra cloud setup</li>
           <li>Founder-license support — we help you wire the riskiest flows first: migrations, force-pushes, deploys, and CI</li>
@@ -835,7 +839,7 @@ __GA_BOOTSTRAP__
       <a href="https://www.linkedin.com/in/igorganapolsky" target="_blank" rel="noopener">LinkedIn</a>
       <a href="/blog">Blog</a>
     </div>
-    <span class="footer-copy">© 2026 Max Smith KDP LLC · MIT License · v1.0.0</span>
+    <span class="footer-copy">© 2026 Max Smith KDP LLC · MIT License · v1.1.0</span>
   </div>
 </footer>
 

package/scripts/export-hf-dataset.js

@@ -0,0 +1,293 @@
+#!/usr/bin/env node
+'use strict';
+
+/**
+ * HuggingFace Dataset Exporter
+ *
+ * Exports ThumbGate agent traces as a HuggingFace-compatible dataset in two formats:
+ *
+ * 1. Agent Traces (traces split) — raw feedback entries with tool calls, signals,
+ *    context, and outcomes. Matches the "share your agent traces" initiative.
+ *
+ * 2. DPO Preferences (preferences split) — chosen/rejected preference pairs
+ *    derived from error→learning memory promotion. Ready for DPO/RLHF training.
+ *
+ * Output: Parquet-compatible JSONL files + dataset_info.json (HF Dataset Card metadata).
+ *
+ * HuggingFace Datasets format:
+ *   dataset_dir/
+ *     dataset_info.json        — metadata, features schema, splits
+ *     traces.jsonl             — agent trace rows
+ *     preferences.jsonl        — DPO preference pair rows
+ */
+
+const fs = require('fs');
+const path = require('path');
+const { resolveFeedbackDir } = require('./feedback-paths');
+const { exportDpoFromMemories } = require('./export-dpo-pairs');
+const { getProvenance } = require('./contextfs');
+
+// ---------------------------------------------------------------------------
+// Helpers
+// ---------------------------------------------------------------------------
+
+function readJSONL(filePath) {
+  if (!fs.existsSync(filePath)) return [];
+  const raw = fs.readFileSync(filePath, 'utf-8').trim();
+  if (!raw) return [];
+  return raw
+    .split('\n')
+    .map((line) => {
+      try { return JSON.parse(line); } catch { return null; }
+    })
+    .filter(Boolean);
+}
+
+function ensureDir(dirPath) {
+  if (!fs.existsSync(dirPath)) {
+    fs.mkdirSync(dirPath, { recursive: true });
+  }
+}
+
+function writeJSONL(filePath, rows) {
+  const content = rows.map((row) => JSON.stringify(row)).join('\n');
+  fs.writeFileSync(filePath, content ? `${content}\n` : '');
+}
+
+// ---------------------------------------------------------------------------
+// PII / path redaction
+// ---------------------------------------------------------------------------
+
+function redactPaths(text) {
+  if (!text || typeof text !== 'string') return text || '';
+  return text
+    .replace(/\/Users\/[^\s/]+/g, '/Users/redacted')
+    .replace(/\/home\/[^\s/]+/g, '/home/redacted')
+    .replace(/C:\\Users\\[^\s\\]+/g, 'C:\\Users\\redacted');
+}
+
+function redactEntry(obj) {
+  if (!obj || typeof obj !== 'object') return obj;
+  const out = {};
+  for (const [key, value] of Object.entries(obj)) {
+    if (typeof value === 'string') {
+      out[key] = redactPaths(value);
+    } else if (Array.isArray(value)) {
+      out[key] = value.map((v) => (typeof v === 'string' ? redactPaths(v) : v));
+    } else {
+      out[key] = value;
+    }
+  }
+  return out;
+}
+
+// ---------------------------------------------------------------------------
+// Trace row builder — converts feedback-log entries to HF trace rows
+// ---------------------------------------------------------------------------
+
+function buildTraceRow(entry, index) {
+  return {
+    trace_id: entry.id || `trace_${index}`,
+    timestamp: entry.timestamp || null,
+    signal: entry.signal || entry.feedback || 'unknown',
+    tool_name: entry.toolName || entry.actionType || 'unknown',
+    context: redactPaths(entry.context || ''),
+    what_worked: redactPaths(entry.whatWorked || ''),
+    what_went_wrong: redactPaths(entry.whatWentWrong || ''),
+    what_to_change: redactPaths(entry.whatToChange || ''),
+    tags: Array.isArray(entry.tags) ? entry.tags : [],
+    failure_type: entry.failureType || null,
+    source: 'thumbgate',
+  };
+}
+
+// ---------------------------------------------------------------------------
+// Preference row builder — converts DPO pairs to HF preference rows
+// ---------------------------------------------------------------------------
+
+function buildPreferenceRow(pair, index) {
+  return {
+    pair_id: `pref_${index}`,
+    prompt: redactPaths(pair.prompt || ''),
+    chosen: redactPaths(pair.chosen || ''),
+    rejected: redactPaths(pair.rejected || ''),
+    match_score: pair.metadata ? pair.metadata.matchScore : null,
+    matched_keys: pair.metadata ? pair.metadata.matchedKeys || [] : [],
+    rubric_delta: pair.metadata && pair.metadata.rubric
+      ? pair.metadata.rubric.weightedDelta
+      : null,
+    source: 'thumbgate',
+  };
+}
+
+// ---------------------------------------------------------------------------
+// Dataset info (HuggingFace Dataset Card metadata)
+// ---------------------------------------------------------------------------
+
+function buildDatasetInfo({ traceCount, preferenceCount, exportedAt }) {
+  return {
+    dataset_info: {
+      description: 'Agent traces and DPO preference pairs from ThumbGate — pre-action gates for AI coding agents. Contains real-world tool call feedback, failure patterns, and learned corrections.',
+      citation: '',
+      homepage: 'https://github.com/IgorGanapolsky/ThumbGate',
+      license: 'MIT',
+      features: {
+        traces: {
+          trace_id: { dtype: 'string' },
+          timestamp: { dtype: 'string' },
+          signal: { dtype: 'string' },
+          tool_name: { dtype: 'string' },
+          context: { dtype: 'string' },
+          what_worked: { dtype: 'string' },
+          what_went_wrong: { dtype: 'string' },
+          what_to_change: { dtype: 'string' },
+          tags: { dtype: 'list', inner: { dtype: 'string' } },
+          failure_type: { dtype: 'string' },
+          source: { dtype: 'string' },
+        },
+        preferences: {
+          pair_id: { dtype: 'string' },
+          prompt: { dtype: 'string' },
+          chosen: { dtype: 'string' },
+          rejected: { dtype: 'string' },
+          match_score: { dtype: 'float32' },
+          matched_keys: { dtype: 'list', inner: { dtype: 'string' } },
+          rubric_delta: { dtype: 'float32' },
+          source: { dtype: 'string' },
+        },
+      },
+      splits: {
+        traces: { num_examples: traceCount },
+        preferences: { num_examples: preferenceCount },
+      },
+    },
+    exported_at: exportedAt,
+    exporter: 'thumbgate/export-hf-dataset',
+    version: '1.0.0',
+  };
+}
+
+// ---------------------------------------------------------------------------
+// Main export function
+// ---------------------------------------------------------------------------
+
+/**
+ * Export ThumbGate data as a HuggingFace-compatible dataset.
+ *
+ * @param {Object} options
+ * @param {string} [options.outputDir] - Directory to write dataset files
+ * @param {string} [options.feedbackDir] - Override feedback data directory
+ * @param {boolean} [options.includeProvenance] - Include provenance events in traces
+ * @returns {Object} Export summary
+ */
+function exportHfDataset(options = {}) {
+  const feedbackDir = options.feedbackDir || resolveFeedbackDir();
+  const outputDir = options.outputDir || path.join(feedbackDir, 'hf-dataset');
+  const includeProvenance = options.includeProvenance !== false;
+
+  ensureDir(outputDir);
+
+  // --- Traces split ---
+  const feedbackLogPath = path.join(feedbackDir, 'feedback-log.jsonl');
+  const feedbackEntries = readJSONL(feedbackLogPath);
+  const traceRows = feedbackEntries.map((entry, i) => buildTraceRow(redactEntry(entry), i));
+
+  // Optionally append provenance events as traces
+  if (includeProvenance) {
+    try {
+      const provenanceEvents = getProvenance(200);
+      for (const evt of provenanceEvents) {
+        traceRows.push({
+          trace_id: evt.id || `prov_${traceRows.length}`,
+          timestamp: evt.timestamp || null,
+          signal: 'provenance',
+          tool_name: evt.type || 'context_assembly',
+          context: redactPaths(JSON.stringify(evt).slice(0, 500)),
+          what_worked: '',
+          what_went_wrong: '',
+          what_to_change: '',
+          tags: ['provenance'],
+          failure_type: null,
+          source: 'thumbgate',
+        });
+      }
+    } catch {
+      // Provenance read failure should not break export
+    }
+  }
+
+  writeJSONL(path.join(outputDir, 'traces.jsonl'), traceRows);
+
+  // --- Preferences split ---
+  const memoryLogPath = path.join(feedbackDir, 'memory-log.jsonl');
+  const memories = readJSONL(memoryLogPath);
+  let preferenceRows = [];
+
+  if (memories.length > 0) {
+    try {
+      const dpoResult = exportDpoFromMemories(memories);
+      preferenceRows = dpoResult.pairs.map((pair, i) => buildPreferenceRow(pair, i));
+    } catch {
+      // DPO export failure should not break the traces export
+    }
+  }
+
+  writeJSONL(path.join(outputDir, 'preferences.jsonl'), preferenceRows);
+
+  // --- Dataset info ---
+  const exportedAt = new Date().toISOString();
+  const info = buildDatasetInfo({
+    traceCount: traceRows.length,
+    preferenceCount: preferenceRows.length,
+    exportedAt,
+  });
+  fs.writeFileSync(
+    path.join(outputDir, 'dataset_info.json'),
+    JSON.stringify(info, null, 2) + '\n',
+  );
+
+  return {
+    outputDir,
+    traceCount: traceRows.length,
+    preferenceCount: preferenceRows.length,
+    files: ['traces.jsonl', 'preferences.jsonl', 'dataset_info.json'],
+    exportedAt,
+  };
+}
+
+// ---------------------------------------------------------------------------
+// CLI
+// ---------------------------------------------------------------------------
+
+function main() {
+  const args = {};
+  process.argv.slice(2).forEach((arg) => {
+    if (!arg.startsWith('--')) return;
+    const [key, ...rest] = arg.slice(2).split('=');
+    args[key] = rest.length ? rest.join('=') : true;
+  });
+
+  const result = exportHfDataset({
+    outputDir: args.output || undefined,
+    includeProvenance: args.provenance !== 'false',
+  });
+
+  console.log(`Exported HuggingFace dataset to ${result.outputDir}`);
+  console.log(`  Traces: ${result.traceCount}`);
+  console.log(`  Preferences: ${result.preferenceCount}`);
+  console.log(`  Files: ${result.files.join(', ')}`);
+}
+
+if (require.main === module) {
+  main();
+}
+
+module.exports = {
+  exportHfDataset,
+  buildTraceRow,
+  buildPreferenceRow,
+  buildDatasetInfo,
+  redactPaths,
+  redactEntry,
+  readJSONL,
+};

package/scripts/tool-registry.js

@@ -399,6 +399,17 @@ const TOOLS = [
       },
     },
   }),
+  destructiveTool({
+    name: 'export_hf_dataset',
+    description: 'Export ThumbGate agent traces and DPO preference pairs as a HuggingFace-compatible dataset. Produces traces.jsonl, preferences.jsonl, and dataset_info.json with PII-redacted paths. Ready for huggingface-cli upload.',
+    inputSchema: {
+      type: 'object',
+      properties: {
+        outputDir: { type: 'string', description: 'Output directory (default: feedback-dir/hf-dataset)' },
+        includeProvenance: { type: 'boolean', description: 'Include provenance events in traces (default: true)' },
+      },
+    },
+  }),
   destructiveTool({
     name: 'export_databricks_bundle',
     description: 'Export ThumbGate logs and proof artifacts as a Databricks-ready analytics bundle',