claude-skills/docs/commands/cs-pulse.md

title	description
/cs-pulse — Slash Command for AI Coding Agents	/cs:pulse <topic> — Multi-source recency research. Grill-me intake (topic / angle / window / scope), then parallel Reddit + HN + Web (1 q/sec per. Slash command for Claude Code, Codex CLI, Gemini CLI.

/cs-pulse

:material-console: Slash Command :material-github: Source

Command: /cs:pulse <topic>

The cs-pulse persona takes the pulse of a topic across Reddit, Hacker News, the open web, and (optionally) X/Twitter — within a configurable recent window — and synthesizes a single coherent briefing.

When to Run

• "What are people saying about X right now?"
• Competitor research with recency flavor
• Trend discovery / tool comparisons / audience sentiment
• Pre-content-creation reconnaissance

The skill ALSO triggers automatically without /cs:pulse when you use trigger phrases:

• "pulse on [topic]"
• "what's happening with [topic]"
• "what are people saying about [topic]"
• "current conversation about [topic]"
• "take the pulse of [topic]"
• "trending: [topic]"
• "find me info on [topic]"

/cs:pulse is the explicit form.

Forcing Intake (2–4 Questions, One at a Time)

Q	Asks	Why
Q1	Topic specificity (1–2 sentences, no vague nouns)	Vague Q1 → vague briefing. Refuses "AI" / "tech" once.
Q2	Angle: trend / sentiment / problems / opportunities / comparison	Dictates which platform's voice weights more in synthesis. Default: trend.
Q3	Time window: 7 / 14 / 30 / 60 / 90 days	Default: 30. 7d = breaking, 90d = sustained shift.
Q4	Platform scope (skip any?)	Asked only when angle suggests some platforms off-target. Default: all.

What You Get

# [TOPIC] — Pulse (Last [N] Days)
*Generated: [DATE] | Angle: [trend|sentiment|problems|opportunities|comparison]*

## TL;DR
[2-3 sentences]

## Reddit
### Top Posts ... ### What Reddit Is Saying

## Hacker News
### Notable Stories ... ### What HN Is Saying

## Web
### Key Sources ... ### What the Web Is Saying

## X/Twitter (if available)
[Or: "Skipped — [reason]"]

## Cross-Platform Patterns
## Key Takeaways
## Content Angles (if applicable)

---
*Audit:* Queries sent: N. Sources received: M. Sources cited: K.

Saved to ${RESEARCH_DIR}/pulse/<topic-slug>-<YYYY-MM-DD>.md AND pasted in chat.

Discipline

• One intake question per turn. Never bundle.
• Refuse vague Q1 once. Push back with examples; deliver with caveat if user won't narrow.
• Parallel Phases 1–3 — Reddit + HN + Web concurrent. Sequential within platform. 1 q/sec.
• Source discipline — cite only session-call results. [Background] for training knowledge, excluded from cited count.
• Three-count tracking — sent / received / cited in audit log.
• Retry once after 3s — then log. 3 consecutive failures across sources → stop.
• Graceful degradation — single source failure → continue with rest. Never fail the whole run on one source.

Workflow

# A. Pre-flight (post-intake)
python ../skills/pulse/scripts/time_window_calculator.py --window 30d
python ../skills/pulse/scripts/topic_slug_generator.py --topic "<topic>" --date $(date +%Y-%m-%d)
python ../skills/pulse/scripts/citation_tracker.py --action start --session NAME

# B. Phases 1–3 (parallel, 1 q/sec per platform)
#    Reddit: search.json sort=top&t=month + sort=new&t=month + top thread comments
#    HN: Algolia stories + comments with timestamp filter
#    Web: 2–3 targeted queries

# C. Phase 4 (optional, sequential): X/Twitter via Grok / X API / browser automation

# D. Synthesis: cross-platform pattern detection

# E. Output: file + chat + audit summary
python ../skills/pulse/scripts/citation_tracker.py --action close --session NAME

Stop Conditions

• All 4 phases complete (or Phase 4 skipped with note) → synthesize + deliver
• 3 consecutive failures across all sources → stop, report what was collected
• User says "stop" → produce partial briefing with what's been collected so far

Anti-Patterns Rejected

• Starting any search before Q1 (topic specificity) commits
• Batching intake questions
• Hardcoded URLs that won't survive API changes (note format, explain may evolve)
• Specific person/brand references
• Tight coupling to one X/Twitter interface
• Missing fallback behavior
• "Just use [specific tool]" without explaining what the tool does
• Citing training knowledge as session results
• Fabricating sources to fill out a section

Related

• Agent: cs-pulse
• Skill: pulse
• Source spec: megaprompts/01-pulse-megaprompt.md (maintainer-local draft spec — gitignored, not in the public repo)
• Sibling research skills (after build): /cs:litreview, /cs:grants, /cs:syllabus, /cs:patent, /cs:dossier, /cs:research (router)

---

Version: 1.0.0 Source: Path-B direct conversion of megaprompts/01-pulse-megaprompt.md