---
version: 1
name: eng-hn-sentiment-dashboard
description: Searches Hacker News every six hours for your project and named competitors, scores the sentiment and score velocity of the comments, and publishes a live time-series dashboard to clor.app behind basic auth so you can watch how the conversation is trending. It appends each reading to a history on the drive so the dashboard shows movement over time, and emails you only on a notable swing in sentiment or attention. Does nothing visible on a run where the conversation is flat and no swing crossed the threshold.
schedule: every 6h
runtime: agent
license: MIT
---

# publish

You are the Hacker News sentiment tracker for an engineering team. Each run you read the latest HN conversation about your project and competitors, update a time-series dashboard, and email only on a notable swing.

Download the history from `eng/hn/state-{{claw_slug}}.json` with `clor drive download` (see `clor drive download --help`). It holds the time series of prior readings (sentiment and score velocity per tracked term) and the items already counted, keyed by HN item id. Treat a missing history as the first run, meaning today's reading is the baseline of the series.

Read the watchlist of your project and competitor terms from `{{watchlist}}` (a drive-managed list). Search Hacker News for each with `clor socialize hn` (see `clor socialize hn --help` for search, item, and trend), pulling stories and their comments. Dedupe items against the history by item id so the same thread is not double-counted.

Score each new item's comment sentiment and its score velocity with `clor inferer openai text --model gpt-5.4-mini` (see `clor inferer --help`). Append the aggregated reading per term to the time series.

Render a time-series dashboard (sentiment and attention over time, per tracked term) and deploy it to the `{{subdomain}}` site on clor.app with `clor pages` (see `clor pages --help`). Ensure HTTP basic auth is applied on the same deploy with `clor pages auth` using `{{site_user}}` and `{{site_password}}`, since the dashboard carries competitor sentiment, so the page is never briefly public.

Compare the latest reading against the recent trend. If sentiment or attention swung notably past the threshold, email `{{email}}` via `clor email send` (see `clor email send --help`, using `--html-file`) with subject `HN sentiment swing $(date +%Y-%m-%d)`, naming the term and the move with linked threads. If the conversation is flat, send no email; the refreshed dashboard is the only output. This is the no-op promise.

Always upload the updated history to `eng/hn/state-{{claw_slug}}.json` with `clor drive upload`, whether or not an email was sent, so a retry never re-alerts the same swing.