CommunityAbandoned· 1.5y★ 3.7by pskill9

Hacker News MCP

Parses news.ycombinator.com for structured story data across top, new, ask, show, and jobs with configurable limits.

View repo

Safety & Trust

Delv Safety Grade: C

Score 58/100 · assessed 2026-04-28

Maintainer40

Permissions85

Supply chain35

Transparency65

Incidents100

This community MCP server scrapes Hacker News to deliver structured story data. It's maintained by a solo developer (pskill9) with limited public profile and sparse repository activity. The permissions model is reasonably safe since it only performs read-only web scraping of a public site with no authentication or sensitive data access. However, the supply chain is weak: there's no npm package, requiring manual clone-and-build installation with npm install and build steps. The code is open source on GitHub with basic documentation explaining the available tools and parameters. No security incidents are known, but the lack of package distribution and single-maintainer model present modest risks. The scraping approach is straightforward HTML parsing without unusual dependencies, which limits attack surface. Suitable for non-critical use cases where you need HN data in Claude, but the manual installation and solo maintenance warrant caution for production deployments.

Lethal Trifecta (prompt-injection exposure)

TWO OF THREE

Private dataNo

Reads secrets, credentials, private files

Untrusted inputYes

Ingests web pages, PRs, issues, emails

External commsYes

Can send data outbound

Comments and submissions are attacker-controlled. Outbound when posting. The HN crowd is generally good but not fully trustable.

Green flags

Read-only access to public data only, no authentication required
Open source with clear documentation of available tools
No environment variables or secrets needed
Straightforward HTML parsing without complex dependencies
No known security incidents or malicious behaviour

Red flags

Solo maintainer with limited public profile and sparse commit history
No npm package distribution, requires manual clone and build
Web scraping could break if HN changes HTML structure
No versioning or release management visible

Permissions requested

Outbound network

Assessed by Delv Editorial using public metadata. Grades are advisory and update as the ecosystem changes. They do not replace your own review of permissions and code before granting an agent access to sensitive systems.

Install

npm install && npm run build

Review

This MCP server scrapes Hacker News and hands you structured JSON for stories across the front page, new, ask, show, and jobs sections. It's a straightforward parser that turns HN's HTML into something you can actually work with in Claude. You get titles, URLs, scores, comment counts, and submission times without touching a browser. I'd reach for this when I'm building daily briefings or tracking specific topics on HN. The configurable limit parameter means you can grab the top 10 stories or pull 50 if you're doing deeper analysis. It works well for asking Claude to summarise trending discussions or find projects related to whatever you're researching. The jobs section is particularly useful if you're monitoring hiring trends in specific technologies. The implementation is clean but basic. It's a web scraper, so it breaks if HN changes their markup. There's no caching, which means every request hits the site fresh. That's fine for occasional use but wasteful if you're polling frequently. The structured output is consistent, which matters more than you'd think when you're chaining this with other tools. One specific workflow: I use it to generate a weekly digest of Ask HN threads about topics I care about. Claude fetches the stories, filters by keywords in the title, then summarises the top comments. Takes about 30 seconds and beats manually scrolling through pages. Who shouldn't bother: if you need real-time HN data or historical archives, this isn't it. It only grabs what's currently visible on each section. There's no search, no user profiles, no threading beyond what's in the initial story data. For those use cases, you want the official HN API via Algolia. This is for people who want current stories in a format Claude can immediately work with, nothing more.

Verdict

Install this if you regularly ask Claude about current tech discussions or want automated HN briefings. Skip it if you need historical data, real-time updates, or anything beyond the five main sections. It does one thing competently and doesn't pretend otherwise.

Good at

Returns clean, structured JSON that Claude can immediately parse and work with.
Covers all five main HN sections including jobs, which most scrapers ignore.
Configurable story limits mean you control the scope of each request.
No API keys or authentication required, just install and run.
Lightweight implementation with minimal dependencies.

Watch out

Breaks if Hacker News changes their HTML structure, as all scrapers do.
No caching means every request hits the site, wasteful for repeated queries.
Limited to currently visible stories, no historical data or search capability.
Manual config required for hosts beyond Claude Desktop despite claimed Cursor support.
No comment threading or user data beyond what appears in story listings.

Use cases

tech trend tracking
news briefings
project discovery
discussion aggregation

Getting started

1. Clone the repo and run `npm install && npm run build` in the project directory. 2. Add the server to your Claude Desktop config file, pointing the command to the built index.js file with node. 3. Restart Claude Desktop and verify the connection by asking Claude to fetch the top 5 stories from Hacker News. 4. Test the limit parameter by requesting different story counts to confirm it's working as expected. 5. Watch out for rate limiting if you're making frequent requests, as this hits HN directly with no caching layer.

Works with

Claude DesktopClaude CodeCursor