Meeting Intel Brief

Meeting Intel Brief

Overview

Research any company + person before a meeting, synthesize the intel, and output a beautiful single-page HTML brief you can open in your browser and reference during the call.

Core principle: Great meetings are won in the prep, not the room. The goal isn't more information — it's the right information, organized by what you need to know first.

Inputs (ask if missing)

Before starting, collect:

1. Who — Person's name + role (if known)
2. Company — Company name and/or website URL (LinkedIn URL also works)
3. Meeting type — Sales / Investor / Job interview / Partnership / Client briefing / Other
4. Your context — Your name, company/product, and what you want to achieve in this meeting

If the user gives a partial request ("prep for my call with Acme tomorrow"), infer what you can and ask only for what's truly missing.

Requirements

• read-website-fast MCP — used by agents to read URLs. If not installed, fall back to WebFetch and WebSearch tools. Install: github.com/pskl/read-website-fast
• superpowers:dispatching-parallel-agents — for running the 3 research agents concurrently. If not available, run sequentially.

Research Pipeline

Run these 3 agents in parallel using superpowers:dispatching-parallel-agents:

Agent 1 — Company Intel

• Read: homepage, About page, Pricing (if B2B), Blog (recent 2-3 posts)
• Search: "[company name]" news site:techcrunch.com OR site:reuters.com (last 6 months)
• Search: [company name] funding OR acquisition OR launch OR layoffs 2025 OR 2026
• Check job postings (what they're investing in right now is the signal)
• Extract: what they do, who they serve, funding stage/size, 1-2 recent milestones

Agent 2 — Person Intel

• Search LinkedIn: [name] [company] site:linkedin.com
• Search public content: "[name]" "[company]" interview OR talk OR article OR podcast
• Look for: career trajectory, what they write/speak about publicly, any shared background
• Note any mutual connections, schools, companies, or interests

Agent 3 — Context Intel

• Load references/meeting-types.md → find the relevant meeting type section
• Search: [company name] competitors and [company name] vs [competitor]
• Search: [company name] [your company/product] (any prior relationship?)
• Identify the 1-2 most relevant "bridges" — things that connect their world to yours

After all 3 complete (sequential):

Synthesize into this JSON structure, then call the script:

{
  "person": {
    "name": "Sarah Chen",
    "role": "VP of Sales",
    "background": "10 years in SaaS sales, previously at Salesforce",
    "cares_about": "Pipeline velocity and rep productivity",
    "hook": "She gave a talk on outbound in 2024 — you can reference it"
  },
  "company": {
    "name": "Acme Corp",
    "url": "https://acme.com",
    "snapshot": ["B2B SaaS, 200-person company", "Series B $40M in Jan 2026", "Serves mid-market retail", "Hiring 10 AEs — signals expansion"]
  },
  "meeting": {
    "type": "sales",
    "date": "2026-04-10",
    "why_it_matters": "They just raised and are scaling fast. You solve exactly the bottleneck they're hiring around."
  },
  "openers": [
    "I saw your January raise — congrats. What's changed in your pipeline process since then?",
    "You spoke at OutboundConf last year about rep ramp time — we're seeing the same problem everywhere.",
    "Your CTO's recent post about tooling consolidation caught my eye — that's exactly what we help with."
  ],
  "questions": [
    "Where's the biggest drop-off in your current pipeline — top, middle, or close?",
    "What does your current stack look like for [relevant category]?",
    "What would a win look like from your side in the next 90 days?"
  ],
  "watch_out": [
    "They just switched CRM providers — may have integration fatigue",
    "Budget cycles reset in Q3 — don't push for immediate commitment"
  ]
}

Meeting type values: sales | investor | interview | partnership | client | other

python3 scripts/generate_brief.py --json data.json --output meeting-brief-sarah-chen-2026-04-10.html
# Windows: start meeting-brief-*.html
# Mac/Linux: open meeting-brief-*.html

Brief Structure

The JSON above maps to these 7 rendered sections:

| # | Section | What Goes Here | |---|---------|----------------| | 1 | Header | Person name + role, company + meeting type badge, date | | 2 | Company Snapshot | 4-5 bullets: what they do, size/stage, customers, 1 recent milestone | | 3 | Person Intel | 2-line career summary, what they care about, any shared hook | | 4 | Why This Meeting Matters | 2-3 sentences: what they want, what you want, where alignment is | | 5 | 5 Conversation Openers | Each tied to something specific and real from the research | | 6 | 3 Questions to Ask | Designed to uncover what you actually need to know | | 7 | Watch Out For | 1-2 sensitivities or friction points to navigate |

Quality Gate

Before delivering the brief, verify all of these:

• [ ] Every conversation opener references something specific and real — a recent news item, article, hire, launch, or stated belief. Not "I see you're growing fast."
• [ ] "Why This Meeting Matters" is specific to THIS person and THIS meeting — not a generic template sentence
• [ ] No section contains placeholder text, "not found", or is left empty
• [ ] HTML opens without errors and all sections render
• [ ] Person intel has at least one hook that connects your background to theirs

If any gate fails: run another research pass on that specific section. Do not deliver incomplete briefs.

When NOT to Use

• Recurring internal team syncs or standups — over-engineering
• Fully informal social catch-ups
• Cases where the meeting is with someone you know extremely well

References

• references/meeting-types.md — What to prioritize by meeting type (investor, sales, interview, partner, client)
• references/research-sources.md — Where to find intel for different types of people and companies

Scripts

• scripts/generate_brief.py — Generates the HTML brief from structured JSON data. Stdlib-only, zero pip installs. Run python3 scripts/generate_brief.py --help for usage.

Source & license

This open-source skill is cataloged on AgentStack and links to its original source — we do not rehost the code.

• Author: FuzulsFriend
• Source: FuzulsFriend/meeting-intel-brief
• License: MIT

Install and usage instructions live in the source repository linked above.