hn-research

name	hn-research
description	Research Hacker News stories, discussions, and community sentiment using web_search and web_extract. Use when you want to find HN threads about a topic, gauge technical community reaction, or find Ask HN / Show HN posts.
version	1.0.0
author	Hermes Agent
metadata	{"hermes":{"tags":["hackernews","hn","community","tech-discourse"],"category":"research"}}

Hacker News Research

Research Hacker News using web_search and web_extract — the same stack used across all Hermes research skills.

Quick Start

# Find HN discussions about a topic
web_search(query="site:news.ycombinator.com TOPIC")

# Find recent Ask HN threads
web_search(query="site:news.ycombinator.com \"Ask HN\" TOPIC")

# Find Show HN posts
web_search(query="site:news.ycombinator.com \"Show HN\" TOPIC")

# Extract a specific thread (story + comments)
web_extract(urls=["https://news.ycombinator.com/item?id=ITEM_ID"])

Research Workflow

Step 1 — Discover threads

web_search(query="site:news.ycombinator.com TOPIC")
web_search(query="site:news.ycombinator.com TOPIC 2026")

Skim the results for:

• Thread titles with high comment counts (shown in search snippets)
• Ask HN threads ("Ask HN: Anyone using X?") — these reveal real practitioner experience
• Show HN posts — these are launches/demos that got community attention

Step 2 — Extract high-signal threads

For each thread worth reading:

web_extract(urls=["https://news.ycombinator.com/item?id=ITEM_ID"])

Firecrawl renders the page and returns the story + top-level comments as markdown. Read the top 10-20 comments for signal.

Step 3 — Broaden if needed

# Search for the topic in HN comments (not just titles)
web_search(query="site:news.ycombinator.com TOPIC criticism OR alternative OR broken")
web_search(query="site:news.ycombinator.com TOPIC vs OR compared OR benchmark")

Step 4 — Check for specific paper/repo discussions

If you're researching an arXiv paper or GitHub repo:

web_search(query="site:news.ycombinator.com \"PAPER TITLE\"")
web_search(query="site:news.ycombinator.com \"github.com/OWNER/REPO\"")

Common Patterns

Goal	Query
Community reaction to a tool	site:news.ycombinator.com TOOL_NAME
Real-world usage reports	site:news.ycombinator.com "Ask HN" TOPIC
Launch discussions	site:news.ycombinator.com "Show HN" TOPIC
Criticism / issues	site:news.ycombinator.com TOPIC criticism OR issues OR problems
Paper discussion	site:news.ycombinator.com "arxiv" TOPIC

Sentiment Signals

When reading threads, look for:

• Top comment tone — sets the frame for the whole discussion
• Vote patterns — high points + high comments = strong interest; high comments + implied low points = controversial
• Recurring complaints — same issue mentioned by multiple commenters = real problem
• Expert identity signals — "I work at X", "we built this" comments carry more weight
• Ask HN answer quality — a question with many detailed responses = practitioners actively using this

Output Format

When reporting HN findings, structure as:

## Hacker News Signals: [Topic]

### Top Threads
1. [Thread title] — [date if visible]
   URL: https://news.ycombinator.com/item?id=...
   Summary: [1-2 sentences on what the community said]

### Community Sentiment
[positive / negative / mixed / polarized — and why]

### Key Themes from Comments
- ...

### Ask HN Insights
[Any Ask HN threads revealing real-world usage or practitioner experience]

Tips

• Search snippets often show comment counts — prioritize threads with 100+ comments
• HN indexes fast; recent launches (days old) often already have threads
• Use web_extract on the item page rather than trying to paginate comments — Firecrawl gets the top fold cleanly
• If a thread is very long (500+ comments), web_extract will get the top comments which are usually the highest quality