feat: add 5 research tools and blocks framework setup

- Add @tpmjs/tools-page-brief for URL content extraction
- Add @tpmjs/tools-compare-pages for cross-source validation
- Add @tpmjs/tools-source-credibility for credibility scoring
- Add @tpmjs/tools-claim-checklist for factual claim extraction
- Add @tpmjs/tools-timeline-from-text for timeline generation
- Move createBlogPost to packages/tools/official/
- Add blocks.yml for Blocks framework validation
- Update pnpm-workspace.yaml to include official tools

.changeset/add-research-tools.md

@@ -0,0 +1,21 @@
+---
+"@tpmjs/tools-page-brief": minor
+"@tpmjs/tools-compare-pages": minor
+"@tpmjs/tools-source-credibility": minor
+"@tpmjs/tools-claim-checklist": minor
+"@tpmjs/tools-timeline-from-text": minor
+---
+
+Add 5 new research tools for AI-powered content analysis
+
+New tools using AI SDK v6 beta (tool() + jsonSchema() pattern):
+
+- **@tpmjs/tools-page-brief**: Fetch URL and extract summary with key points and claims needing citations
+- **@tpmjs/tools-compare-pages**: Compare two URLs for agreements, conflicts, and unique points
+- **@tpmjs/tools-source-credibility**: Calculate heuristic credibility score based on domain signals
+- **@tpmjs/tools-claim-checklist**: Extract checkable factual claims from text
+- **@tpmjs/tools-timeline-from-text**: Extract dated events and return normalized timeline
+
+All tools are stub implementations ready for full logic implementation.
+
+Also moved @tpmjs/createblogpost to packages/tools/official/ directory and set up blocks.yml for Blocks framework validation.

docs/TODO.md

@@ -130,6 +130,27 @@ GET  /api/users/[username]/collections   - Get user's public collections
 
 ---
 
+## Benchmark Comparison Page
+
+### Overview
+- [ ] Create `/benchmarks` page to compare tool performance
+- [ ] Display execution time comparisons across similar tools
+- [ ] Show token usage efficiency metrics
+- [ ] Compare success rates and reliability
+
+### Features
+- [ ] Side-by-side tool comparison UI
+- [ ] Historical performance trends (charts)
+- [ ] Filter by category to compare relevant tools
+- [ ] Export benchmark data as JSON/CSV
+
+### Data Collection
+- [ ] Track execution metrics from playground simulations
+- [ ] Aggregate anonymous performance data
+- [ ] Calculate percentile rankings within categories
+
+---
+
 ## Notes
 
 - Consider rate limiting on ratings to prevent abuse

package.json

@@ -40,6 +40,8 @@
   },
   "devDependencies": {
     "@biomejs/biome": "^1.9.4",
+    "@blocksai/cli": "^0.2.1",
+    "@blocksai/validators": "^1.1.1",
     "@changesets/cli": "^2.27.10",
     "@total-typescript/ts-reset": "^0.6.1",
     "@types/node": "^22.10.2",

packages/tools/official/.gitignore

@@ -0,0 +1,3 @@
+# Blocks CLI runtime files
+.blocks/
+.env

packages/tools/official/BLOCKS_DOMAIN_VALIDATOR_BUG.md

@@ -0,0 +1,168 @@
+# Blocks Domain Validator Bug Report
+
+## Status: FIXED
+
+The fix has been applied to `/Users/ajaxdavis/repos/blocks/packages/ai/src/provider.ts`.
+
+---
+
+## Summary
+
+The domain validator in `@blocksai/validators` was failing with an OpenAI structured output schema error when validating any block. The schema and shape.ts validators worked correctly.
+
+## Error Message
+
+```
+⚠ [domain] AI validation failed: Invalid schema for response_format 'response':
+In context=('properties', 'issues', 'items'), 'required' is required to be supplied
+and to be an array including every key in properties. Missing 'file'.
+```
+
+## Environment
+
+- `@blocksai/cli`: latest (installed via npx)
+- `@blocksai/validators`: latest
+- OpenAI API: Using `OPENAI_API_KEY` from .env
+- Node.js: v22.x
+- OS: macOS
+
+## Reproduction Steps
+
+1. Create a valid `blocks.yml` with any block definition
+2. Create the corresponding TypeScript file with proper exports
+3. Run `npx blocks run --all`
+
+## Example blocks.yml
+
+```yaml
+name: "tpmjs-official-tools"
+root: "."
+
+philosophy:
+  - "Tools must be pure functions with no side effects"
+
+domain:
+  entities:
+    url:
+      fields: [href, domain, protocol, path]
+
+  signals:
+    credibility:
+      description: "Trustworthiness of a source"
+
+  measures:
+    valid_output:
+      constraints:
+        - "Must return structured object matching interface"
+
+blocks:
+  domain_rules:
+    - id: pure_function
+      description: "Tool must be deterministic with no side effects"
+
+  research.pageBrief:
+    description: "Fetch URL and extract content"
+    path: "page-brief"
+    inputs:
+      - name: url
+        type: string
+    outputs:
+      - name: brief
+        type: PageBrief
+        measures: [valid_output]
+
+validators:
+  - schema
+  - shape.ts
+  - domain
+```
+
+## Observed Behavior
+
+```
+📦 Validating: research.pageBrief
+  ✓ schema ok
+  ✓ shape.ts ok
+-   Running domain...
+  ⚠ [domain] AI validation failed: Invalid schema for response_format 'response':
+     In context=('properties', 'issues', 'items'), 'required' is required to be
+     supplied and to be an array including every key in properties. Missing 'file'.
+
+  ⚠️  Block "research.pageBrief" has warnings
+```
+
+## Root Cause Analysis
+
+The error message indicates that the domain validator is constructing an OpenAI structured output request with a JSON schema that's missing required fields. Specifically:
+
+1. OpenAI's structured output feature requires that all properties in an object schema must be listed in the `required` array
+2. The domain validator's internal response schema has a `file` property in the `issues.items` object
+3. This `file` property is not included in the corresponding `required` array
+
+This is an internal schema construction issue within the domain validator, not related to user-provided blocks.yml configuration.
+
+## Expected Behavior
+
+The domain validator should:
+1. Construct a valid JSON schema for OpenAI's structured output API
+2. Ensure all properties are listed in `required` arrays
+3. Successfully analyze the block against the domain rules and philosophy
+
+## Workaround
+
+Currently, the schema and shape.ts validators work correctly. The domain validator can be skipped by removing `domain` from the validators list in blocks.yml:
+
+```yaml
+validators:
+  - schema
+  - shape.ts
+  # - domain  # Disabled due to bug
+```
+
+## Fix Applied
+
+**File:** `/Users/ajaxdavis/repos/blocks/packages/ai/src/provider.ts` (lines 163-173)
+
+**Before (broken):**
+```typescript
+const schema = z.object({
+  isValid: z.boolean(),
+  issues: z.array(
+    z.object({
+      message: z.string(),
+      severity: z.enum(["error", "warning"]),
+      file: z.string().optional(),  // ← Problem: .optional() excludes from required
+    })
+  ),
+  summary: z.string().optional().describe("..."),
+});
+```
+
+**After (fixed):**
+```typescript
+const schema = z.object({
+  isValid: z.boolean(),
+  issues: z.array(
+    z.object({
+      message: z.string().describe("Description of the issue found"),
+      severity: z.enum(["error", "warning"]).describe("Severity of the issue"),
+      file: z.string().describe("File path where the issue was found, or empty string if not file-specific"),
+    })
+  ),
+  summary: z.string().describe("Brief summary of why the block passed or failed validation"),
+});
+```
+
+**Root Cause:** OpenAI's structured output requires ALL properties to be in the JSON schema's `required` array. When Zod converts `.optional()` to JSON schema, it omits that property from `required`, causing OpenAI to reject the schema.
+
+**Solution:** Remove `.optional()` and `.default()` modifiers. Make all fields required strings. The AI will return an empty string for file-agnostic issues.
+
+## Impact
+
+- **Severity**: Medium - domain validation is non-functional
+- **Affected**: All blocks using the domain validator
+- **Workaround available**: Yes - disable domain validator
+
+## Additional Context
+
+This was tested with 6 different blocks, all showing the same error. The error is consistent and reproducible regardless of block configuration.

packages/tools/official/blocks.yml

@@ -0,0 +1,132 @@
+name: "tpmjs-official-tools"
+root: "."
+
+philosophy:
+  - "Tools must be pure functions with no side effects"
+  - "Each tool solves one specific problem well"
+  - "Tools are publishable to npm as @tpmjs/* packages"
+  - "Use AI SDK v6 tool() + jsonSchema() pattern"
+
+domain:
+  entities:
+    url:
+      fields: [href, domain, protocol, path]
+    text:
+      fields: [content, sentences, wordCount]
+    claim:
+      fields: [statement, needsCitation, evidence]
+    timeline_event:
+      fields: [date, description, confidence]
+
+  signals:
+    credibility:
+      description: "Trustworthiness of a source"
+      extraction_hint: "Look for author, date, citations, HTTPS"
+    readability:
+      description: "How easy content is to understand"
+      extraction_hint: "Check sentence length, jargon, structure"
+
+  measures:
+    valid_output:
+      constraints:
+        - "Must return structured object matching interface"
+        - "Must not throw unhandled errors"
+    npm_publishable:
+      constraints:
+        - "Must have valid package.json with tpmjs field"
+        - "Must export tool as named and default export"
+
+blocks:
+  domain_rules:
+    - id: pure_function
+      description: "Tool must be deterministic with no side effects"
+    - id: ai_sdk_pattern
+      description: "Must use AI SDK v6 tool() + jsonSchema()"
+    - id: npm_ready
+      description: "Must be publishable to npm with complete metadata"
+
+  adapter.createBlogPost:
+    description: "Creates structured blog posts with frontmatter and metadata"
+    path: "createBlogPost"
+    inputs:
+      - name: title
+        type: string
+      - name: author
+        type: string
+      - name: content
+        type: string
+      - name: tags
+        type: string[]
+        optional: true
+      - name: format
+        type: "'markdown' | 'mdx'"
+        optional: true
+    outputs:
+      - name: blogPost
+        type: BlogPost
+        measures: [valid_output, npm_publishable]
+
+  research.pageBrief:
+    description: "Fetch URL, extract main content, return summary with key points and claims needing citations"
+    path: "page-brief"
+    inputs:
+      - name: url
+        type: string
+    outputs:
+      - name: brief
+        type: PageBrief
+        measures: [valid_output]
+
+  research.comparePages:
+    description: "Compare two URLs for agreements, conflicts, and unique points"
+    path: "compare-pages"
+    inputs:
+      - name: urlA
+        type: string
+      - name: urlB
+        type: string
+    outputs:
+      - name: comparison
+        type: PageComparison
+        measures: [valid_output]
+
+  research.sourceCredibility:
+    description: "Heuristic credibility score based on domain signals, author presence, citations"
+    path: "source-credibility"
+    inputs:
+      - name: url
+        type: string
+      - name: html
+        type: string
+        optional: true
+    outputs:
+      - name: credibility
+        type: CredibilityScore
+        measures: [valid_output]
+
+  research.claimChecklist:
+    description: "Extract checkable factual claims from text, mark what needs citations"
+    path: "claim-checklist"
+    inputs:
+      - name: text
+        type: string
+    outputs:
+      - name: checklist
+        type: ClaimChecklist
+        measures: [valid_output]
+
+  research.timelineFromText:
+    description: "Extract dated events from text, return normalized timeline with confidence"
+    path: "timeline-from-text"
+    inputs:
+      - name: text
+        type: string
+    outputs:
+      - name: timeline
+        type: Timeline
+        measures: [valid_output]
+
+validators:
+  - schema
+  - shape.ts
+  - domain

packages/tools/official/claim-checklist/package.json

@@ -0,0 +1,61 @@
+{
+  "name": "@tpmjs/tools-claim-checklist",
+  "version": "0.1.0",
+  "description": "Extract checkable factual claims from text and identify what needs citations",
+  "type": "module",
+  "keywords": ["tpmjs", "research", "ai", "claims", "fact-checking", "citations"],
+  "exports": {
+    ".": {
+      "types": "./dist/index.d.ts",
+      "default": "./dist/index.js"
+    }
+  },
+  "files": ["dist"],
+  "scripts": {
+    "build": "tsup",
+    "dev": "tsup --watch",
+    "type-check": "tsc --noEmit",
+    "clean": "rm -rf dist .turbo"
+  },
+  "devDependencies": {
+    "@tpmjs/tsconfig": "workspace:*",
+    "tsup": "^8.3.5",
+    "typescript": "^5.9.3"
+  },
+  "publishConfig": {
+    "access": "public"
+  },
+  "repository": {
+    "type": "git",
+    "url": "https://github.com/anthropics/tpmjs.git",
+    "directory": "packages/tools/official/claim-checklist"
+  },
+  "homepage": "https://tpmjs.com",
+  "license": "MIT",
+  "tpmjs": {
+    "category": "research",
+    "frameworks": ["vercel-ai"],
+    "tools": [
+      {
+        "name": "claimChecklistTool",
+        "description": "Extract checkable factual claims from text and mark what needs citations",
+        "parameters": [
+          {
+            "name": "text",
+            "type": "string",
+            "description": "The text to analyze for factual claims",
+            "required": true
+          }
+        ],
+        "returns": {
+          "type": "ClaimChecklist",
+          "description": "Object with claims array and summary statistics"
+        }
+      }
+    ]
+  },
+  "dependencies": {
+    "ai": "6.0.0-beta.124",
+    "sbd": "^1.0.19"
+  }
+}

packages/tools/official/claim-checklist/src/index.ts

@@ -0,0 +1,56 @@
+import { jsonSchema, tool } from 'ai';
+
+export interface ClaimChecklist {
+  originalText: string;
+  claims: Array<{
+    claim: string;
+    needsCitation: boolean;
+    evidenceType: 'statistic' | 'fact' | 'quote' | 'opinion' | 'common-knowledge';
+    suggestedEvidence: string;
+    sentenceIndex: number;
+  }>;
+  summary: {
+    totalClaims: number;
+    needingCitation: number;
+    verified: number;
+  };
+}
+
+type ClaimChecklistInput = {
+  text: string;
+};
+
+export const claimChecklistTool = tool({
+  description:
+    'Extract checkable factual claims from text, mark what needs citations, and suggest what evidence would support each claim',
+  inputSchema: jsonSchema<ClaimChecklistInput>({
+    type: 'object',
+    properties: {
+      text: {
+        type: 'string',
+        description: 'The text to analyze for factual claims',
+      },
+    },
+    required: ['text'],
+    additionalProperties: false,
+  }),
+  async execute({ text }): Promise<ClaimChecklist> {
+    // TODO: Implement with:
+    // 1. Split text into sentences with sbd
+    // 2. Rule-based claim extraction (numbers, dates, proper nouns)
+    // 3. Classify claim types
+    // 4. Determine if citation needed based on type
+
+    return {
+      originalText: text,
+      claims: [],
+      summary: {
+        totalClaims: 0,
+        needingCitation: 0,
+        verified: 0,
+      },
+    };
+  },
+});
+
+export default claimChecklistTool;

packages/tools/official/compare-pages/package.json

@@ -0,0 +1,67 @@
+{
+  "name": "@tpmjs/tools-compare-pages",
+  "version": "0.1.0",
+  "description": "Compare two URLs for agreements, conflicts, and unique points for cross-source validation",
+  "type": "module",
+  "keywords": ["tpmjs", "research", "ai", "comparison", "fact-checking"],
+  "exports": {
+    ".": {
+      "types": "./dist/index.d.ts",
+      "default": "./dist/index.js"
+    }
+  },
+  "files": ["dist"],
+  "scripts": {
+    "build": "tsup",
+    "dev": "tsup --watch",
+    "type-check": "tsc --noEmit",
+    "clean": "rm -rf dist .turbo"
+  },
+  "devDependencies": {
+    "@tpmjs/tsconfig": "workspace:*",
+    "tsup": "^8.3.5",
+    "typescript": "^5.9.3"
+  },
+  "publishConfig": {
+    "access": "public"
+  },
+  "repository": {
+    "type": "git",
+    "url": "https://github.com/anthropics/tpmjs.git",
+    "directory": "packages/tools/official/compare-pages"
+  },
+  "homepage": "https://tpmjs.com",
+  "license": "MIT",
+  "tpmjs": {
+    "category": "research",
+    "frameworks": ["vercel-ai"],
+    "tools": [
+      {
+        "name": "comparePagesTool",
+        "description": "Compare two URLs for agreements, conflicts, and unique points",
+        "parameters": [
+          {
+            "name": "urlA",
+            "type": "string",
+            "description": "First URL to compare",
+            "required": true
+          },
+          {
+            "name": "urlB",
+            "type": "string",
+            "description": "Second URL to compare",
+            "required": true
+          }
+        ],
+        "returns": {
+          "type": "PageComparison",
+          "description": "Object with agreements, conflicts, uniqueToA, and uniqueToB arrays"
+        }
+      }
+    ]
+  },
+  "dependencies": {
+    "ai": "6.0.0-beta.124",
+    "natural": "^8.0.0"
+  }
+}

packages/tools/official/compare-pages/src/index.ts

@@ -0,0 +1,57 @@
+import { jsonSchema, tool } from 'ai';
+
+export interface PageComparison {
+  urlA: string;
+  urlB: string;
+  agreements: string[];
+  conflicts: Array<{
+    topic: string;
+    pageAPosition: string;
+    pageBPosition: string;
+  }>;
+  uniqueToA: string[];
+  uniqueToB: string[];
+}
+
+type ComparePagesInput = {
+  urlA: string;
+  urlB: string;
+};
+
+export const comparePagesTool = tool({
+  description:
+    'Compare two URLs for agreements, conflicts, and unique points to help with cross-source validation',
+  inputSchema: jsonSchema<ComparePagesInput>({
+    type: 'object',
+    properties: {
+      urlA: {
+        type: 'string',
+        description: 'First URL to compare',
+      },
+      urlB: {
+        type: 'string',
+        description: 'Second URL to compare',
+      },
+    },
+    required: ['urlA', 'urlB'],
+    additionalProperties: false,
+  }),
+  async execute({ urlA, urlB }): Promise<PageComparison> {
+    // TODO: Implement with:
+    // 1. Call pageBriefTool on both URLs
+    // 2. Extract key points from both
+    // 3. Use TF-IDF (natural) for semantic matching
+    // 4. Identify agreements, conflicts, and unique points
+
+    return {
+      urlA,
+      urlB,
+      agreements: [],
+      conflicts: [],
+      uniqueToA: [],
+      uniqueToB: [],
+    };
+  },
+});
+
+export default comparePagesTool;

packages/tools/official/compare-pages/tsconfig.json

@@ -0,0 +1,11 @@
+{
+  "extends": "@tpmjs/tsconfig/base.json",
+  "compilerOptions": {
+    "outDir": "dist",
+    "rootDir": "src",
+    "incremental": false,
+    "composite": false
+  },
+  "include": ["src"],
+  "exclude": ["node_modules", "dist"]
+}

packages/tools/official/compare-pages/tsup.config.ts

@@ -0,0 +1,10 @@
+import { defineConfig } from 'tsup';
+
+export default defineConfig({
+  entry: ['src/index.ts'],
+  format: ['esm'],
+  dts: true,
+  clean: true,
+  treeshake: true,
+  splitting: false,
+});

packages/tools/official/createBlogPost/package.json

@@ -28,7 +28,7 @@
   "repository": {
     "type": "git",
     "url": "https://github.com/ajaxdavis/tpmjs.git",
-    "directory": "packages/tools/createBlogPost"
+    "directory": "packages/tools/official/createBlogPost"
   },
   "homepage": "https://tpmjs.com",
   "license": "MIT",

packages/tools/official/createBlogPost/tsconfig.json

@@ -0,0 +1,11 @@
+{
+  "extends": "@tpmjs/tsconfig/base.json",
+  "compilerOptions": {
+    "outDir": "dist",
+    "rootDir": "src",
+    "incremental": false,
+    "composite": false
+  },
+  "include": ["src"],
+  "exclude": ["node_modules", "dist"]
+}

packages/tools/official/createBlogPost/tsup.config.ts

@@ -0,0 +1,10 @@
+import { defineConfig } from 'tsup';
+
+export default defineConfig({
+  entry: ['src/index.ts'],
+  format: ['esm'],
+  dts: true,
+  clean: true,
+  treeshake: true,
+  splitting: false,
+});

packages/tools/official/page-brief/package.json

@@ -0,0 +1,63 @@
+{
+  "name": "@tpmjs/tools-page-brief",
+  "version": "0.1.0",
+  "description": "Fetch a URL, extract main content, and return a brief with summary, key points, and claims needing citations",
+  "type": "module",
+  "keywords": ["tpmjs", "research", "ai", "readability", "web-scraping"],
+  "exports": {
+    ".": {
+      "types": "./dist/index.d.ts",
+      "default": "./dist/index.js"
+    }
+  },
+  "files": ["dist"],
+  "scripts": {
+    "build": "tsup",
+    "dev": "tsup --watch",
+    "type-check": "tsc --noEmit",
+    "clean": "rm -rf dist .turbo"
+  },
+  "devDependencies": {
+    "@tpmjs/tsconfig": "workspace:*",
+    "tsup": "^8.3.5",
+    "typescript": "^5.9.3"
+  },
+  "publishConfig": {
+    "access": "public"
+  },
+  "repository": {
+    "type": "git",
+    "url": "https://github.com/anthropics/tpmjs.git",
+    "directory": "packages/tools/official/page-brief"
+  },
+  "homepage": "https://tpmjs.com",
+  "license": "MIT",
+  "tpmjs": {
+    "category": "research",
+    "frameworks": ["vercel-ai"],
+    "tools": [
+      {
+        "name": "pageBriefTool",
+        "description": "Fetch a URL, extract main content, and return a brief with summary, key points, and claims needing citations",
+        "parameters": [
+          {
+            "name": "url",
+            "type": "string",
+            "description": "The URL to fetch and analyze",
+            "required": true
+          }
+        ],
+        "returns": {
+          "type": "PageBrief",
+          "description": "Object with url, title, summary, keyPoints array, and claimsNeedingCitation array"
+        }
+      }
+    ]
+  },
+  "dependencies": {
+    "ai": "6.0.0-beta.124",
+    "@mozilla/readability": "^0.5.0",
+    "jsdom": "^26.0.0",
+    "sbd": "^1.0.19"
+  }
+}

packages/tools/official/page-brief/src/index.ts

@@ -0,0 +1,50 @@
+import { jsonSchema, tool } from 'ai';
+
+export interface PageBrief {
+  url: string;
+  title: string;
+  summary: string;
+  keyPoints: string[];
+  claimsNeedingCitation: Array<{
+    claim: string;
+    suggestedEvidence: string;
+  }>;
+}
+
+type PageBriefInput = {
+  url: string;
+};
+
+export const pageBriefTool = tool({
+  description:
+    'Fetch a URL, extract main content using readability, and return a brief with summary, key points, and claims that need citations',
+  inputSchema: jsonSchema<PageBriefInput>({
+    type: 'object',
+    properties: {
+      url: {
+        type: 'string',
+        description: 'The URL to fetch and analyze',
+      },
+    },
+    required: ['url'],
+    additionalProperties: false,
+  }),
+  async execute({ url }): Promise<PageBrief> {
+    // TODO: Implement with fetch + @mozilla/readability + jsdom + sbd
+    // 1. Fetch the URL
+    // 2. Parse HTML with jsdom
+    // 3. Extract main content with Readability
+    // 4. Split into sentences with sbd
+    // 5. Identify claims that need citations
+
+    return {
+      url,
+      title: 'Not implemented',
+      summary: 'This is a stub implementation. Real implementation will use @mozilla/readability.',
+      keyPoints: [],
+      claimsNeedingCitation: [],
+    };
+  },
+});
+
+export default pageBriefTool;

packages/tools/official/page-brief/tsconfig.json

@@ -0,0 +1,11 @@
+{
+  "extends": "@tpmjs/tsconfig/base.json",
+  "compilerOptions": {
+    "outDir": "dist",
+    "rootDir": "src",
+    "incremental": false,
+    "composite": false
+  },
+  "include": ["src"],
+  "exclude": ["node_modules", "dist"]
+}

packages/tools/official/page-brief/tsup.config.ts

@@ -0,0 +1,10 @@
+import { defineConfig } from 'tsup';
+
+export default defineConfig({
+  entry: ['src/index.ts'],
+  format: ['esm'],
+  dts: true,
+  clean: true,
+  treeshake: true,
+  splitting: false,
+});

packages/tools/official/source-credibility/package.json

@@ -0,0 +1,68 @@
+{
+  "name": "@tpmjs/tools-source-credibility",
+  "version": "0.1.0",
+  "description": "Calculate heuristic credibility score based on domain signals, author presence, and citations",
+  "type": "module",
+  "keywords": ["tpmjs", "research", "ai", "credibility", "fact-checking"],
+  "exports": {
+    ".": {
+      "types": "./dist/index.d.ts",
+      "default": "./dist/index.js"
+    }
+  },
+  "files": ["dist"],
+  "scripts": {
+    "build": "tsup",
+    "dev": "tsup --watch",
+    "type-check": "tsc --noEmit",
+    "clean": "rm -rf dist .turbo"
+  },
+  "devDependencies": {
+    "@tpmjs/tsconfig": "workspace:*",
+    "tsup": "^8.3.5",
+    "typescript": "^5.9.3"
+  },
+  "publishConfig": {
+    "access": "public"
+  },
+  "repository": {
+    "type": "git",
+    "url": "https://github.com/anthropics/tpmjs.git",
+    "directory": "packages/tools/official/source-credibility"
+  },
+  "homepage": "https://tpmjs.com",
+  "license": "MIT",
+  "tpmjs": {
+    "category": "research",
+    "frameworks": ["vercel-ai"],
+    "tools": [
+      {
+        "name": "sourceCredibilityTool",
+        "description": "Calculate heuristic credibility score based on domain signals",
+        "parameters": [
+          {
+            "name": "url",
+            "type": "string",
+            "description": "URL to analyze for credibility",
+            "required": true
+          },
+          {
+            "name": "html",
+            "type": "string",
+            "description": "Optional pre-fetched HTML content",
+            "required": false
+          }
+        ],
+        "returns": {
+          "type": "CredibilityScore",
+          "description": "Object with score (0-1), signals breakdown, and detailed analysis"
+        }
+      }
+    ]
+  },
+  "dependencies": {
+    "ai": "6.0.0-beta.124",
+    "tldts": "^6.1.0",
+    "cheerio": "^1.0.0"
+  }
+}

packages/tools/official/source-credibility/src/index.ts

@@ -0,0 +1,75 @@
+import { jsonSchema, tool } from 'ai';
+
+export interface CredibilityScore {
+  url: string;
+  score: number; // 0.0 to 1.0
+  signals: {
+    hasHttps: boolean;
+    hasAuthor: boolean;
+    hasPublishDate: boolean;
+    hasCitations: boolean;
+    domainAge?: string;
+    isKnownSource: boolean;
+  };
+  breakdown: Array<{
+    signal: string;
+    weight: number;
+    present: boolean;
+  }>;
+}
+
+type SourceCredibilityInput = {
+  url: string;
+  html?: string;
+};
+
+export const sourceCredibilityTool = tool({
+  description:
+    'Calculate a heuristic credibility score based on domain signals, author presence, date, citations density, and HTTPS',
+  inputSchema: jsonSchema<SourceCredibilityInput>({
+    type: 'object',
+    properties: {
+      url: {
+        type: 'string',
+        description: 'URL to analyze for credibility',
+      },
+      html: {
+        type: 'string',
+        description: 'Optional: pre-fetched HTML content to analyze',
+      },
+    },
+    required: ['url'],
+    additionalProperties: false,
+  }),
+  async execute({ url, html: _html }): Promise<CredibilityScore> {
+    // TODO: Implement with:
+    // 1. Parse URL with tldts for domain info
+    // 2. If no HTML provided, fetch the page
+    // 3. Parse HTML with cheerio for meta signals
+    // 4. Check for author, date, citations
+    // 5. Calculate weighted score
+
+    const isHttps = url.startsWith('https://');
+
+    return {
+      url,
+      score: 0.5, // Stub score
+      signals: {
+        hasHttps: isHttps,
+        hasAuthor: false,
+        hasPublishDate: false,
+        hasCitations: false,
+        isKnownSource: false,
+      },
+      breakdown: [
+        { signal: 'https', weight: 0.1, present: isHttps },
+        { signal: 'author', weight: 0.2, present: false },
+        { signal: 'publishDate', weight: 0.2, present: false },
+        { signal: 'citations', weight: 0.3, present: false },
+        { signal: 'knownSource', weight: 0.2, present: false },
+      ],
+    };
+  },
+});
+
+export default sourceCredibilityTool;

packages/tools/official/source-credibility/tsconfig.json

@@ -0,0 +1,11 @@
+{
+  "extends": "@tpmjs/tsconfig/base.json",
+  "compilerOptions": {
+    "outDir": "dist",
+    "rootDir": "src",
+    "incremental": false,
+    "composite": false
+  },
+  "include": ["src"],
+  "exclude": ["node_modules", "dist"]
+}

packages/tools/official/source-credibility/tsup.config.ts

@@ -0,0 +1,10 @@
+import { defineConfig } from 'tsup';
+
+export default defineConfig({
+  entry: ['src/index.ts'],
+  format: ['esm'],
+  dts: true,
+  clean: true,
+  treeshake: true,
+  splitting: false,
+});

packages/tools/official/timeline-from-text/package.json

@@ -0,0 +1,62 @@
+{
+  "name": "@tpmjs/tools-timeline-from-text",
+  "version": "0.1.0",
+  "description": "Extract dated events from text and return a normalized timeline with confidence scores",
+  "type": "module",
+  "keywords": ["tpmjs", "research", "ai", "timeline", "dates", "events"],
+  "exports": {
+    ".": {
+      "types": "./dist/index.d.ts",
+      "default": "./dist/index.js"
+    }
+  },
+  "files": ["dist"],
+  "scripts": {
+    "build": "tsup",
+    "dev": "tsup --watch",
+    "type-check": "tsc --noEmit",
+    "clean": "rm -rf dist .turbo"
+  },
+  "devDependencies": {
+    "@tpmjs/tsconfig": "workspace:*",
+    "tsup": "^8.3.5",
+    "typescript": "^5.9.3"
+  },
+  "publishConfig": {
+    "access": "public"
+  },
+  "repository": {
+    "type": "git",
+    "url": "https://github.com/anthropics/tpmjs.git",
+    "directory": "packages/tools/official/timeline-from-text"
+  },
+  "homepage": "https://tpmjs.com",
+  "license": "MIT",
+  "tpmjs": {
+    "category": "research",
+    "frameworks": ["vercel-ai"],
+    "tools": [
+      {
+        "name": "timelineFromTextTool",
+        "description": "Extract dated events from text and return normalized timeline",
+        "parameters": [
+          {
+            "name": "text",
+            "type": "string",
+            "description": "The text to extract timeline events from",
+            "required": true
+          }
+        ],
+        "returns": {
+          "type": "Timeline",
+          "description": "Object with events array and optional date range"
+        }
+      }
+    ]
+  },
+  "dependencies": {
+    "ai": "6.0.0-beta.124",
+    "chrono-node": "^2.7.0",
+    "sbd": "^1.0.19"
+  }
+}

packages/tools/official/timeline-from-text/src/index.ts

@@ -0,0 +1,50 @@
+import { jsonSchema, tool } from 'ai';
+
+export interface Timeline {
+  originalText: string;
+  events: Array<{
+    date: string; // ISO format
+    description: string;
+    confidence: number; // 0.0 to 1.0
+    originalMention: string;
+  }>;
+  dateRange?: {
+    earliest: string;
+    latest: string;
+  };
+}
+
+type TimelineFromTextInput = {
+  text: string;
+};
+
+export const timelineFromTextTool = tool({
+  description:
+    'Extract dated events from text and return a normalized timeline with confidence scores per event',
+  inputSchema: jsonSchema<TimelineFromTextInput>({
+    type: 'object',
+    properties: {
+      text: {
+        type: 'string',
+        description: 'The text to extract timeline events from',
+      },
+    },
+    required: ['text'],
+    additionalProperties: false,
+  }),
+  async execute({ text }): Promise<Timeline> {
+    // TODO: Implement with:
+    // 1. Use chrono-node to parse dates from text
+    // 2. Extract sentence context around each date
+    // 3. Normalize dates to ISO format
+    // 4. Assign confidence based on date specificity
+
+    return {
+      originalText: text,
+      events: [],
+      dateRange: undefined,
+    };
+  },
+});
+
+export default timelineFromTextTool;

packages/tools/official/timeline-from-text/tsconfig.json

@@ -0,0 +1,11 @@
+{
+  "extends": "@tpmjs/tsconfig/base.json",
+  "compilerOptions": {
+    "outDir": "dist",
+    "rootDir": "src",
+    "incremental": false,
+    "composite": false
+  },
+  "include": ["src"],
+  "exclude": ["node_modules", "dist"]
+}

packages/tools/official/timeline-from-text/tsup.config.ts

@@ -0,0 +1,10 @@
+import { defineConfig } from 'tsup';
+
+export default defineConfig({
+  entry: ['src/index.ts'],
+  format: ['esm'],
+  dts: true,
+  clean: true,
+  treeshake: true,
+  splitting: false,
+});

pnpm-workspace.yaml

@@ -5,3 +5,4 @@ packages:
   - 'packages/config/tailwind'
   - 'packages/config/tsconfig'
   - 'packages/tools/*'
+  - 'packages/tools/official/*'