Pulse — Multi-Source Recency Research — Research

Pulse is a multi-source recency research skill that takes the pulse of a topic across Reddit, Hacker News, the open web, and optionally X/Twitter within a configurable time window (default 30 days). It synthesizes findings into a single briefing with citations, engagement metrics, and cross-platform pattern analysis, capturing the current conversation rather than canonical references.

research claude-code

📂 View skill on GitHub ↗ 🚀 How to use

✨ What it does

Forces intake on topic specificity, angle, time window, and platform scope before searching.
Searches Reddit, Hacker News, and the open web in parallel, with optional X/Twitter.
Tracks queries sent, sources received, and sources cited via a citation tracker.
Synthesizes cross-platform patterns: consensus, controversy, pain points, excitement, emerging trends, and gaps.
Outputs a structured markdown briefing with TL;DR, per-platform sections, and key takeaways.
Includes error handling for rate limits, empty results, and unavailable platforms.

🎯 When to use it

When you need to know what people are currently saying about a topic across multiple platforms.
For competitor research, trend discovery, tool comparisons, or audience sentiment analysis with a recency focus.
When you want a synthesized briefing with citations and engagement metrics from Reddit, HN, and the web.
When you need to identify consensus, controversy, pain points, and emerging trends on a topic.

🚀 How to use

Trigger: /cs:grill-with-docs

Trigger Pulse with phrases like 'pulse on [topic]', 'what's happening with [topic]', or 'take the pulse of [topic]'. The skill will ask up to 4 forcing questions about topic specificity, angle, time window, and platform scope. Provide a specific topic and answer the questions to proceed. Example prompts:

pulse on self-hosted LLM deployment for small teams
what are people saying about Claude Code adoption?

📄 Output: A markdown briefing saved to ${RESEARCH_DIR}/pulse/<topic-slug>-<YYYY-MM-DD>.md and pasted in chat, containing TL;DR, per-platform summaries, cross-platform patterns, and key takeaways.

📦 Add this skill to Claude Code

# 1. Get the skills repo
git clone --depth 1 https://github.com/alirezarezvani/claude-skills /tmp/claude-skills

# 2. Copy this skill into your project (or ~/.claude/skills for all projects)
mkdir -p .claude/skills
cp -r /tmp/claude-skills/research/pulse/skills/pulse .claude/skills/pulse

Skill source: research/pulse/skills/pulse/SKILL.md

⚠️ Good to know

Requires a specific topic; vague topics like 'AI' will be pushed back on. X/Twitter data is optional and may be skipped if no interface is available. Rate limits apply per platform (1 q/sec).

❓ FAQ

What happens if I give a vague topic like 'AI'?

The skill will push back once, asking you to narrow the topic to an angle like adoption, safety, or regulation. If you still don't narrow it, it will proceed with a 'vague topic' caveat and deliver a survey-level briefing.

Can I skip X/Twitter?

Yes, during intake you can specify platforms to skip. If X/Twitter is not available (no browser automation, Grok, or API), the skill will include a section header with 'Skipped — [reason]' and not pretend to have data.

How does the skill handle rate limits?

It throttles to 1 query per second per platform. On failure, it waits 3 seconds and retries once. After 3 consecutive failures across all sources, it stops and returns what was collected with an explicit 'stopped early' note.

🤖 Overview, features, install steps and FAQ were generated from the project's SKILL.md on Sep 4, 2026. Always check the original source before running commands.