Pulse Articles

Goal

Research trending topics and write informative explainers for a technical audience. Gather crowd conversation from Reddit, X, YouTube, Hacker News, GitHub, and Polymarket over the last 30 days, verify key facts against primary sources, and publish a structured briefing that explains what's happening and what it means.

Use Cases

—Write an explainer for a trending tool or technique you've seen discussed
—Aggregate latest developments and verified facts about a new release or framework
—Catalog community sentiment and signal from multiple platforms
—Understand emerging topics with a grounded view of durable value and practical limits

Prerequisites

•Authenticated CLI
•At least one research engine configured
•Python 3.12+ (for the last30days skill)
•The last30days skill installed (auto-discovered or set RADAR_PULSE_SKILL_DIR)

Commands

radar pulse <topic>

Research a trending topic. Gathers last-30-days conversation, verifies the latest developments, writes an explainer, publishes it, and posts one personal LinkedIn teaser.

radar pulse <topic> --post-linkedin

Backward-compatible alias. Personal LinkedIn posting is already automatic after a successful publication.

radar pulse <topic> --post-linkedin-company <org-id>

Override the personal target and post the published explainer only to the selected LinkedIn company page.

radar topics

Shows trending Hacker News, GitHub, and Reddit topics, then prompts to select multiple for pulse articles.

radar content update <pulse-slug>

Re-research an existing pulse article using the stored topic. Runs the full pulse pipeline fresh.

radar content cover <pulse-slug>

Generate or refresh the Higgsfield cover image for a pulse article (when Higgsfield CLI is authenticated).

Environment Variables

RADAR_PULSE_SKILL_DIR

Override last30days skill discovery. Points to the directory containing last30days.py.

How It Works

**Stage 1: Gather crowd conversation** — The local last30days skill searches Reddit, X, YouTube, TikTok, Hacker News, GitHub, Polymarket, and the web for the topic, collecting quotes, engagement stats, and sentiment.

**Stage 2: Verify facts** — Your selected research engine checks every factual claim against primary sources (official repos, docs, homepages) and distinguishes verified from inferred.

**Stage 3: Write explainer** — The engine writes a structured informative article: what it is, why it's trending, what's changed, what it means, and an assessment of durable value.

**Stage 4: Review and publish** — Judge/editor review for clarity and flow, then publish to the catalog as sourceType: pulse.

Example Session

$ radar pulse "cursor agent skills"

Research engine: Codex CLI

Locating last30days skill…
✓ Found at ~/.claude/plugins/cache/last30days-skill/last30days/3.6.1

  ⚙ Research — gather crowd conversation (last30days)
  ⚙ Verify facts against primary sources (Codex CLI)
  ⚙ Write informative explainer (Codex CLI)
  ⚙ Judge/editor review (Codex CLI)
  ⚙ Generate cover image (Higgsfield)
  ⚙ Publish to shared catalog

✓ Added to the shared catalog.
✓ Evidence saved to .radar/pulse-runs/cursor-agent-skills-20260812/

SUMMARY
  Cursor Agent Skills · Development Workflows
  What's new in agent skills for Cursor IDE, why it matters now, and where to learn more.
  radar show cursor-agent-skills