freshcontext-mcp 0.3.14 → 0.3.16

package/SESSION_SAVE_V5.md

@@ -1,121 +0,0 @@
-# FreshContext — Session Save V5
-**Date:** 2026-03-27
-**npm:** freshcontext-mcp@0.3.13
-**Tools:** 19 live
-
----
-
-## What Was Done This Session
-
-- npm tokens renewed: freshcontext-publish + NPM_TOKEN (both renewed, NPM_TOKEN updated in GitHub secrets)
-- GitHub granular token renewed
-- v0.3.13 pushed — extract_gebiz added (Singapore GeBIZ procurement via data.gov.sg)
-- README fully rewritten — 19 tools, company landscape section with PLTR demo, all unique adapters documented
-- GovTech Singapore follow-up sent (delivered on commitment: GeBIZ tool built and live)
-- Palantir follow-up sent (references PLTR company landscape report, $1.1B contracts, Rule of 40 127%)
-- 10 follow-up drafts created in Gmail (PatSnap, Mistral, HF, Klarna, SAP, Moonshot, MiniMax, Apify, Cloudflare, Sea)
-- PLTR company landscape report confirmed working — full intelligence output documented
-- Apify rebuild needed for v0.3.13 (manual trigger required)
-
----
-
-## Current Tool Count: 19
-
-**Standard (11):** extract_github, extract_hackernews, extract_scholar, extract_arxiv,
-extract_reddit, extract_yc, extract_producthunt, search_repos, package_trends,
-extract_finance, search_jobs
-
-**Composite landscapes (4):**
-- extract_landscape — YC + GitHub + HN + Reddit + Product Hunt + npm
-- extract_gov_landscape — govcontracts + HN + GitHub + changelog
-- extract_finance_landscape — finance + HN + Reddit + GitHub + changelog
-- extract_company_landscape — SEC + govcontracts + GDELT + changelog + finance
-
-**Unique — not in any other MCP server (4 + GeBIZ):**
-- extract_changelog — release history from any repo/package/site
-- extract_govcontracts — US federal contract awards (USASpending.gov)
-- extract_sec_filings — 8-K material event disclosures (SEC EDGAR)
-- extract_gdelt — global news events (GDELT Project, 100+ languages)
-- extract_gebiz — Singapore Government procurement (data.gov.sg) ← NEW
-
----
-
-## Outreach Status
-
-**Active threads:**
-- GovTech Singapore — delivered GeBIZ tool, awaiting response
-- Palantir — follow-up sent with PLTR company landscape report
-
-**Follow-up drafts sitting in Gmail (10):**
-- PatSnap — contact@patsnap.com
-- Mistral AI — contact@mistral.ai
-- Hugging Face — api-enterprise@huggingface.co
-- Klarna — partnerships@klarna.com
-- SAP Startups — startups@sap.com
-- Moonshot AI — support@moonshot.cn
-- MiniMax — contact@minimax.io
-- Apify / Jan — jan@apify.com
-- Cloudflare Startups — startups@cloudflare.com
-- Sea / Shopee — ir@sea.com
-
-**Bounced — need correct addresses (9):**
-- Revolut — bd@revolut.com + press@revolut.com both dead
-- Zalando — partnerships@zalando.de + tech@zalando.de both dead
-- Celonis — partnerships@celonis.com dead
-- Grab — partnerships@grab.com + developer@grab.com both dead
-- Sea Limited — partnerships@sea.com dead (ir@sea.com delivered but wrong team)
-- Zhipu AI — bd@zhipuai.cn + contact@zhipuai.cn both dead
-- MiniMax — bd@minimaxi.com dead (contact@minimax.io delivered)
-- Moonshot AI — business@moonshot.cn dead (support@moonshot.cn delivered)
-- Apollo — hello@apollo.io dead
-
-**New targets identified (not yet contacted):**
-- IMDA Singapore (Infocomm Media Development Authority)
-- Australian Digital Transformation Agency
-- UK Government Digital Service (GDS)
-- LangChain / LangSmith
-- LlamaIndex
-- CrewAI
-- Vercel AI SDK
-- FactSet
-- Morningstar
-
-**Correct addresses to find for bounced:**
-- Revolut → try partnerships@revolut.com
-- Grab → try business@grab.com
-- Celonis → try hello@celonis.com
-- Apollo.io → try partnerships@apollo.io
-- Zhipu AI → LinkedIn outreach (email dead)
-
----
-
-## Pending Items
-
-- Send 10 follow-up drafts from Gmail Drafts folder
-- Trigger Apify rebuild for v0.3.13
-- Find correct addresses for 9 bounced companies
-- Contact new targets: IMDA, GDS, LangChain, LlamaIndex, CrewAI, FactSet
-- GKG upgrade for extract_gdelt (tone scores, goldstein scale) — deferred
-- Agnost AI analytics integration — sign up at app.agnost.ai, one line in server.ts
-- Synthesis endpoint (/briefing/now) — needs ANTHROPIC_KEY + $5 credits
-
----
-
-## Demo Assets
-1. intelligence-report.html — Anthropic/OpenAI/Palantir government intelligence report
-2. PLTR company landscape — Q4 2025 earnings, $1.1B contracts, Rule of 40 127%
-
-Both are best-in-class product demos. Use these in outreach.
-
----
-
-## Resume Prompt
-"I'm building freshcontext-mcp — 19 tools live at v0.3.13. Last session: built extract_gebiz
-(Singapore GeBIZ), sent GovTech and Palantir follow-ups, drafted 10 follow-up emails.
-Next: send the 10 drafts, fix bounced email addresses, contact new targets.
-See SESSION_SAVE_V5.md."
-
----
-
-*"The work isn't gone. It's just waiting to be continued."*
-*— Prince Gabriel, Grootfontein, Namibia 🇳🇦*

package/USAGE.md

@@ -1,294 +0,0 @@
-# FreshContext — Usage Guide
-> *13 tools. No API keys. Every result timestamped.*
-
----
-
-## Quick Start
-
-Add FreshContext to Claude Desktop (no install required):
-
-```json
-{
-  "mcpServers": {
-    "freshcontext": {
-      "command": "npx",
-      "args": ["-y", "mcp-remote", "https://freshcontext-mcp.gimmanuel73.workers.dev/mcp"]
-    }
-  }
-}
-```
-
-Restart Claude. Done. All 13 tools are now available.
-
----
-
-## The 13 Tools
-
-| Tool | What it does | Input |
-|---|---|---|
-| `extract_github` | README, stars, forks, last commit | GitHub URL |
-| `extract_hackernews` | Stories + scores with timestamps | HN URL or search query |
-| `extract_scholar` | Research papers ranked by year | Google Scholar URL |
-| `extract_arxiv` | arXiv papers via official API | Search query or arXiv URL |
-| `extract_reddit` | Subreddit posts + community sentiment | Subreddit URL |
-| `extract_yc` | YC company listings by keyword | Topic keyword |
-| `extract_producthunt` | Recent launches by topic | Topic or PH URL |
-| `search_repos` | GitHub repos ranked by stars + recency | Search query |
-| `package_trends` | npm/PyPI metadata + download stats | Package name(s) |
-| `extract_finance` | Live stock data — price, P/E, 52w range | Ticker symbol(s) |
-| `extract_landscape` | 6 sources in one call | Topic keyword |
-| `extract_changelog` | Update history from any product/repo | GitHub URL, npm name, or website URL |
-| `extract_govcontracts` | US federal contract awards | Company name, keyword, or NAICS code |
-
----
-
-## Command Reference
-
-Copy and paste these directly into Claude.
-
----
-
-### Competitive Research
-
-```
-Use extract_landscape with topic "AI data freshness"
-```
-*Full picture: who's funded, what's trending, what repos exist, what packages are moving.*
-
-```
-Use search_repos with query "mcp server typescript"
-```
-*GitHub repos ranked by stars. See who's already building in your space.*
-
-```
-Use extract_yc with query "developer infrastructure"
-```
-*Which YC companies are funded in your category.*
-
-```
-Use extract_producthunt with url "developer tools"
-```
-*What launched this week in your category.*
-
----
-
-### Market Intelligence
-
-```
-Use extract_hackernews with url "https://hn.algolia.com/?q=model+context+protocol+2026"
-```
-*What the developer community is saying about a topic right now.*
-
-```
-Use extract_reddit with url "https://www.reddit.com/r/MachineLearning/"
-```
-*Subreddit sentiment — what problems people are complaining about or excited about.*
-
-```
-Use extract_finance with url "NVDA,MSFT,GOOGL"
-```
-*Live stock data for competitive set. Know the market before a pitch.*
-
-```
-Use extract_govcontracts with url "AI infrastructure"
-```
-*Which companies are winning federal AI contracts. High-signal GTM data.*
-
-```
-Use extract_govcontracts with url "Palantir"
-```
-*How much government business a specific company is doing — and when.*
-
----
-
-### Job Market Intelligence
-
-```
-Use search_jobs with query "AI engineer" location "remote" max_age_days 7
-```
-*Fresh job listings only. Every result has a freshness badge — never apply to a closed role again.*
-
-```
-Use search_jobs with query "typescript developer" remote_only true
-```
-*Remote-only filtered results with publish dates.*
-
-```
-Use extract_govcontracts with url "541511"
-```
-*NAICS 541511 = Custom Computer Programming Services. See who's hiring via contract awards.*
-
----
-
-### Developer Ecosystem Research
-
-```
-Use package_trends with url "mcp,wrangler,anthropic"
-```
-*npm download trends for multiple packages. See what's gaining traction.*
-
-```
-Use extract_changelog with url "https://github.com/anthropics/anthropic-sdk-python"
-```
-*Latest releases with dates. Know exactly when a dependency shipped a change.*
-
-```
-Use extract_changelog with url "freshcontext-mcp"
-```
-*npm package changelog — version history and release cadence.*
-
-```
-Use extract_changelog with url "https://linear.app"
-```
-*Auto-discovers the changelog page and extracts update history.*
-
-```
-Use search_repos with query "cloudflare workers MCP 2026"
-```
-*What's being built on a platform right now.*
-
----
-
-### Academic & Research Intelligence
-
-```
-Use extract_arxiv with url "https://arxiv.org/search/?query=model+context+protocol&searchtype=all"
-```
-*Latest papers on a topic with submission dates.*
-
-```
-Use extract_scholar with url "https://scholar.google.com/scholar?q=AI+agent+grounding"
-```
-*Google Scholar results ranked by recency.*
-
----
-
-### Full Research Workflows
-
-#### Validate a startup idea
-```
-1. Use extract_landscape with topic "[your idea]"
-2. Use extract_yc with query "[your idea]"
-3. Use extract_govcontracts with url "[your idea keyword]"
-4. Use extract_hackernews with url "https://hn.algolia.com/?q=[your+idea]"
-```
-*Tells you: who's funded, what's trending, is there government demand, what do developers think.*
-
-#### Research a company before a meeting
-```
-1. Use extract_github with url "https://github.com/[company]"
-2. Use extract_changelog with url "https://github.com/[company]/[main-repo]"
-3. Use extract_govcontracts with url "[company name]"
-4. Use extract_finance with url "[ticker]"
-```
-*Tells you: how active is their engineering team, what did they ship recently, do they have government contracts, what's the stock doing.*
-
-#### Find your next job
-```
-1. Use search_jobs with query "[your role]" max_age_days 7
-2. Use extract_govcontracts with url "[target company]"
-3. Use extract_changelog with url "https://github.com/[target-company]/[repo]"
-```
-*Fresh listings only + proof the company is growing + proof they're actively shipping.*
-
-#### Track a competitor
-```
-1. Use extract_changelog with url "https://[competitor].com"
-2. Use extract_github with url "https://github.com/[competitor]/[repo]"
-3. Use package_trends with url "[competitor-npm-package]"
-4. Use extract_hackernews with url "https://hn.algolia.com/?q=[competitor+name]"
-```
-*When did they last ship? Is their repo active? Are downloads growing? What is the community saying?*
-
----
-
-## Who This Is For
-
-### Jobseekers
-The tool FreshContext was built for. Stop applying to roles that closed months ago. Every job listing comes with a freshness badge — green means posted this week, red means posted 60+ days ago. Filter by `max_age_days` to only see roles posted recently.
-
-**Key tools:** `search_jobs`, `extract_govcontracts`, `extract_changelog`
-
----
-
-### Founders & Product Teams
-Validate ideas, track competitors, monitor your ecosystem. `extract_landscape` gives you a full competitive picture in one call — YC funding, GitHub activity, HN sentiment, Product Hunt launches, npm traction, all timestamped.
-
-**Key tools:** `extract_landscape`, `extract_yc`, `extract_changelog`, `extract_hackernews`
-
----
-
-### GTM & Sales Teams
-Government contracts are buying intent signals. A company that just won a $5M DoD contract is hiring, spending, and building. FreshContext is the only MCP server with access to USASpending.gov contract data.
-
-**Key tools:** `extract_govcontracts`, `extract_landscape`, `extract_finance`
-
----
-
-### Investors & Analysts
-Track portfolio companies, find signals before they're obvious, monitor competitor funding. `extract_govcontracts` shows government contract pipeline. `extract_changelog` shows engineering velocity. `extract_finance` gives live market data.
-
-**Key tools:** `extract_govcontracts`, `extract_landscape`, `extract_finance`, `extract_changelog`
-
----
-
-### Researchers & Academics
-Fresh papers from arXiv and Google Scholar, ranked by date. Never cite a paper without knowing exactly when it was published and retrieved.
-
-**Key tools:** `extract_arxiv`, `extract_scholar`, `extract_hackernews`
-
----
-
-### Developers & DevRel Teams
-Monitor your dependencies, track ecosystem traction, see what's getting attention in the community. `extract_changelog` works on any GitHub repo, npm package, or website that has a changelog.
-
-**Key tools:** `extract_changelog`, `package_trends`, `search_repos`, `extract_github`
-
----
-
-## Understanding the FreshContext Envelope
-
-Every result is wrapped like this:
-
-```
-[FRESHCONTEXT]
-Source: https://github.com/owner/repo
-Published: 2026-03-10
-Retrieved: 2026-03-17T09:19:00Z
-Confidence: high
----
-... content ...
-[/FRESHCONTEXT]
-```
-
-**Confidence levels:**
-- `high` — date came from a structured API field. Reliable.
-- `medium` — date inferred from page signals or URL patterns. Likely correct.
-- `low` — no date signal found. Treat with caution.
-
-The gap between `Published` and `Retrieved` is your staleness indicator. If a job was published 90 days ago and retrieved today, the role is likely filled. If a GitHub repo was last pushed 2 years ago, the project may be abandoned.
-
----
-
-## Freshness Score (coming in v0.4.0)
-
-A numeric score from 0–100 will be added to every result:
-
-```
-freshness_score = max(0, 100 - (days_since_retrieved × decay_rate))
-```
-
-Different data types decay at different rates — financial data goes stale faster than academic papers. You'll be able to filter results by `freshness_score > 70` directly in your prompts.
-
----
-
-## The Standard
-
-FreshContext implements the **FreshContext Specification v1.0** — an open standard for data freshness envelopes. See `FRESHCONTEXT_SPEC.md` for the full specification.
-
-Any tool or agent that wraps its retrieved data in the FreshContext envelope is **FreshContext-compatible**.
-
----
-
-*Built by Prince Gabriel — Grootfontein, Namibia 🇳🇦*
-*"The work isn't gone. It's just waiting to be continued."*

package/add-cache.cjs

@@ -1,86 +0,0 @@
-const fs = require('fs');
-let content = fs.readFileSync('./worker/src/worker.ts', 'utf8');
-
-// Tools that need withCache wrapping (github already done)
-// For each: find the async handler start and add withCache open
-// Then find the corresponding catch line and add the closing
-
-const tools = [
-  { name: 'extract_hackernews',  adapter: 'hackernews',    input: 'url' },
-  { name: 'extract_scholar',     adapter: 'scholar',       input: 'url' },
-  { name: 'extract_yc',          adapter: 'yc',            input: 'url' },
-  { name: 'search_repos',        adapter: 'reposearch',    input: 'query' },
-  { name: 'package_trends',      adapter: 'packagetrends', input: 'packages' },
-  { name: 'extract_reddit',      adapter: 'reddit',        input: 'url' },
-  { name: 'extract_producthunt', adapter: 'producthunt',   input: 'url' },
-  { name: 'extract_finance',     adapter: 'finance',       input: 'url' },
-  { name: 'search_jobs',         adapter: 'jobs',          input: 'query' },
-  { name: 'extract_landscape',   adapter: 'landscape',     input: 'topic' },
-];
-
-// We'll work line-by-line for precision
-const lines = content.split('\n');
-const result = [];
-let i = 0;
-
-while (i < lines.length) {
-  const line = lines[i];
-
-  // Check if this line closes a tool handler that needs wrapping
-  // Pattern: "    } catch (err: any) { return ..." followed by "  });"
-  // We add "    }); // end withCache" before the closing  });
-  if (
-    line.match(/^\s+\} catch \(err: any\) \{ return \{ content/) &&
-    lines[i + 1] && lines[i + 1].trim() === '});' &&
-    !(lines[i + 1].includes('end withCache'))
-  ) {
-    result.push(line);
-    result.push('    }); // end withCache');
-    i++;
-    continue;
-  }
-
-  result.push(line);
-  i++;
-}
-
-content = result.join('\n');
-
-// Now add the withCache OPEN for each tool
-// Pattern to find: the }, async ({ INPUT }) => {\n    try {
-// for each tool, but only if not already wrapped
-
-for (const tool of tools) {
-  // Build a specific marker string that is unique per tool
-  // Look for the line: "  }, async ({ url/query/etc }) => {"
-  // followed by next non-empty content being "    try {"
-  // We replace "    try {" with "    return withCache(...); try {"
-  
-  // Find the tool's registerTool call, then find the first "try {" after it
-  const toolMarker = `server.registerTool("${tool.name}"`;
-  const idx = content.indexOf(toolMarker);
-  if (idx === -1) { console.log('Tool not found: ' + tool.name); continue; }
-
-  // From that position, find the first "    try {" that is NOT already preceded by withCache
-  const afterTool = content.slice(idx);
-  const tryIdx = afterTool.indexOf('\n    try {');
-  if (tryIdx === -1) { console.log('No try block for: ' + tool.name); continue; }
-
-  // Check if withCache is already there (look 200 chars before the try)
-  const beforeTry = afterTool.slice(Math.max(0, tryIdx - 200), tryIdx);
-  if (beforeTry.includes('withCache')) {
-    console.log('Already wrapped: ' + tool.name);
-    continue;
-  }
-
-  // Insert the withCache line before "    try {"
-  const absoluteIdx = idx + tryIdx;
-  content = content.slice(0, absoluteIdx) +
-    `\n    return withCache("${tool.adapter}", ${tool.input}, env.CACHE, async () => {` +
-    content.slice(absoluteIdx);
-
-  console.log('Wrapped: ' + tool.name);
-}
-
-fs.writeFileSync('./worker/src/worker.ts', content, 'utf8');
-console.log('\nAll done. Lines: ' + content.split('\n').length);

package/dataset_schema.json

@@ -1,41 +0,0 @@
-{
-  "actorSpecification": 1,
-  "fields": [
-    {
-      "fieldId": "adapter",
-      "fieldType": "String",
-      "title": "Adapter",
-      "description": "The source adapter used to retrieve the data (e.g. github, hackernews, reddit, yc, scholar)."
-    },
-    {
-      "fieldId": "source_url",
-      "fieldType": "String",
-      "title": "Source URL",
-      "description": "The URL of the original source the data was retrieved from."
-    },
-    {
-      "fieldId": "content",
-      "fieldType": "String",
-      "title": "Content",
-      "description": "The retrieved content from the source, truncated to max_length characters."
-    },
-    {
-      "fieldId": "retrieved_at",
-      "fieldType": "String",
-      "title": "Retrieved at",
-      "description": "ISO 8601 timestamp of when FreshContext fetched this data. Always reflects the actual retrieval time."
-    },
-    {
-      "fieldId": "content_date",
-      "fieldType": "String",
-      "title": "Content date",
-      "description": "Best estimate of when the original content was published. Null if unknown."
-    },
-    {
-      "fieldId": "freshness_confidence",
-      "fieldType": "String",
-      "title": "Freshness confidence",
-      "description": "Confidence level of the content_date estimate. One of: high (from structured API/metadata), medium (inferred from page signals), low (estimated or unknown)."
-    }
-  ]
-}

package/input_schema.json

@@ -1,48 +0,0 @@
-{
-  "title": "FreshContext Input",
-  "type": "object",
-  "schemaVersion": 1,
-  "properties": {
-    "tool": {
-      "title": "Tool",
-      "type": "string",
-      "description": "The FreshContext tool to run. See README for full descriptions.",
-      "enum": [
-        "extract_hackernews",
-        "extract_github",
-        "extract_govcontracts",
-        "extract_sec_filings",
-        "extract_gdelt",
-        "extract_gebiz",
-        "extract_finance",
-        "extract_changelog",
-        "extract_arxiv",
-        "extract_scholar",
-        "extract_reddit",
-        "extract_yc",
-        "extract_producthunt",
-        "search_repos",
-        "package_trends",
-        "search_jobs"
-      ],
-      "default": "extract_hackernews",
-      "editor": "select"
-    },
-    "url": {
-      "title": "URL or Query",
-      "type": "string",
-      "description": "The main input for the tool. For extract_github: full GitHub URL. For extract_hackernews: HN URL or search URL. For extract_govcontracts/extract_sec_filings/extract_gdelt/extract_gebiz/extract_finance: company name, keyword, or ticker. For search_repos/package_trends/search_jobs: search keyword.",
-      "editor": "textfield"
-    },
-    "max_length": {
-      "title": "Max content length",
-      "type": "integer",
-      "description": "Maximum characters returned. Default: 8000.",
-      "default": 8000,
-      "minimum": 500,
-      "maximum": 20000,
-      "editor": "number"
-    }
-  },
-  "required": ["tool"]
-}