GREENFIELD v4 — account brief · earned-right enforced100% free · MIT licensed · runs locallyBest in Codex, Claude Code, Cursor, Gemini CLI, Antigravity, ReplitWeb chat works too — with more manual contextGREENFIELD v4 — account brief · earned-right enforced100% free · MIT licensed · runs locallyBest in Codex, Claude Code, Cursor, Gemini CLI, Antigravity, ReplitWeb chat works too — with more manual context
100% free · MIT license · runs locally

Break Open
Greenfield
Accounts.

Find your way into every account in your territory. Greenfield is a structured AI skill that runs a 14-part research process inside Claude Code or Cursor — working every angle (triggers, execs, tech stack, warm paths, earned-right insight) until it finds the wedge.

* Best in CLI tools with browsing or connected data, including Claude Code, Codex, Gemini CLI, Antigravity, Cursor, Replit; web chat still works, but you will need to provide more account context manually.

Visual field guide

See how the skill works.

Explore the whole process, or take it one step at a time.

Download PNG ↓
Greenfield New Logo workflow diagram. All steps are available in the text guide below.
Read the complete step-by-step guide
  1. Account identity

    Confirm the legal entity, domain and business unit.

  2. Trigger radar

    Find dated changes and test why they matter now.

  3. Connection co-pilot

    Use an optional LinkedIn export or guided searches.

  4. Company context

    Adapt research for startup, growth or enterprise.

  5. Internal intelligence

    Capture prior activity, colleagues and existing access.

  6. Website and product

    Understand customers, business model and priorities.

  7. Hiring and growth

    Read roles and hiring patterns as hypotheses.

  8. Market and financials

    Connect pressures to revenue, cost, risk or capital.

  9. Leadership deep dive

    Map individual remits, priorities and public evidence.

  10. Executive narrative

    Choose a route and test the earned right to engage.

  11. Warm path activation

    Rank direct, partner, customer and community routes.

  12. Technical landscape

    Check documented systems and competitive context.

  13. Industry language

    Go one level deeper into how the buyer works.

  14. Persona messaging

    Translate one relevant angle for each stakeholder.

Output: Account Brief: company snapshot • why now • priority personas • warm paths • account wedge • executive narrative • first messages • intelligence gaps. 30-day execution plan: actions, owners, timing and trigger-based adjustments.

Audit claims and access before drafting. Keep verified facts, inferences and unknowns separate.

Greenfield New Logo
Greenfield New Logo full-size field guide

Scroll to inspect the image. Press Escape to close.

01 / What it is

Your AI should interrogate the account, interrogate you, and earn the route in.

Before

Twelve tabs. A 10-K you'll skim. A doc no one reads.

You guess the org chart, copy-paste fragments into notes, and send the same generic outreach every other rep already sent. The account stays cold.

During

14 parts. The model does the work. The gates do the thinking.

Greenfield forces verification, pressure-tests assumptions, and refuses executive outreach until the Earned Right Test passes. Unknowns stay visible as [UNVERIFIED].

After

A War Room Brief with the buyer, the trigger, the path, and the opening move.

You leave with a ranked route in, an earned-right narrative, targeted drafts, a warm intro brief, and a 30-day execution plan. Not a loose note doc.

02 / The 14-part interrogation

Every angle worked, in order — so nothing obvious gets missed.

00Setup

Account Identity Disambiguation

Confirm the exact company, HQ, industry, and what you are selling before any research starts. Wrong account, wrong run.

0.5Urgency

Trigger Radar

Lock in urgency before anything else. If the account has no real pressure, do not pretend the timing is there.

CPNetwork

LinkedIn Connections Co-Pilot

Always offered after Trigger Radar. If you share your LinkedIn export, Greenfield maps direct hits, second-degree paths, connectors, and validation targets before any startup, YC, or enterprise branching.

PSInternal

Pre-Stage Company Context

Only after the LinkedIn decision is handled, branch by your company's actual assets: founders, investors, customers, alumni, partnerships, and product footprint before external research expands.

01Internal

Internal Intelligence

Sweep CRM, Slack, partner overlap, exec relationships, and prior company contact before you search outside.

02External

Company Website & Product Analysis

Become their customer for 20 minutes. Product experience, compliance clues, stack signals, and the real theme the company is running.

03External

Hiring Signals & Growth Intelligence

Where they're hiring is where they're investing. Job patterns expose the problem list faster than generic research.

04Market

Market & Financial Intelligence

Use earnings, funding, investor pressure, board context, and market timing to sharpen the wedge.

05People

Leadership Deep Dive

Profiles, podcasts, interviews, and language analysis to understand how the real decision-makers see the business.

5.5People

Executive Narrative & Access Strategy

Decide whether to target CEO, CTO, CFO, CRO, CMO, CPO, or not go C-suite yet. The Earned Right Test runs here.

06Access

Warm Path Activation

Rank every intro route before cold outreach: direct links, change agents, customers, investors, partners, alumni, and internal relationships.

07Technical

Technical Landscape & Competitive Intelligence

Use the stack, competitive setup, partner ecosystem, and renewal timing to sharpen the technical angle.

08Language

Industry Language & One Level Deeper

Apply the one-level-deeper rule so the messaging sounds like operator insight instead of generic category talk.

09Execute

Persona-Driven Messaging

Turn the research into different hooks for executives, technical buyers, and champions, with conditional drafts when exec outreach is not earned.

03 / Sample War Room Brief

The full rep-facing output, not a stripped-down summary.

Real public-source research, not placeholder copy. This is still an illustrative selling scenario, so product context is generic and some stakeholder lanes are shown as roles rather than invented names.
Greenfield War Room Brief

Anthropic

Frontier AI / enterprise AI platform · San Francisco, CA · Private company · hypergrowth

8.8
Why Now / 10
Run04-18 · public-source run
StatusDONE_WITH_CONCERNS
Verification7.9/10
Earned RightPASSED at partner / enterprise leader level
Warm paths exhaustedNO
Recommendation

Work the partner-and-enterprise route now. CEO path withheld.

Snapshot

Anthropic is not just selling model access anymore. The public signals now point to a fast-expanding enterprise machine: over 1,000 business customers spending more than $1 million annualized, a $100 million Claude Partner Network, large strategic partnerships like Snowflake, and growing infrastructure commitments across AWS, Google, Broadcom, and Microsoft. This is the kind of account where the surface story is model quality, but the real buying motion is partner-led production deployment.

Trigger Radar
HIGH

Demand and infrastructure scaled sharply in April 2026

Anthropic said run-rate revenue surpassed $30 billion, the number of business customers spending over $1 million annualized exceeded 1,000, and it signed for multiple gigawatts of next-generation TPU capacity.

Anthropic + Google/Broadcom announcement · Apr 6, 2026
HIGH

Claude Partner Network launched with $100 million behind it

Anthropic is funding certifications, sales enablement, technical support, and market development so partners can move enterprises from proof of concept to production.

Anthropic Partner Network announcement · Mar 12, 2026
HIGH

Enterprise GTM is getting deeper, not broader

The Snowflake expansion came with a $200 million agreement and a joint GTM motion focused on deploying AI agents in large enterprises.

Snowflake partnership announcement · Dec 3, 2025
MED

Cloud and compliance footprint widened

Anthropic says Claude is available on AWS, Google Cloud, and Microsoft Azure, while government-compliance approvals have expanded through Bedrock and Vertex AI.

Anthropic cloud / compliance announcements · 2025–2026
Account Wedge
Core pain
Partner-led enterprise deployment is scaling faster than field standardization.
Primary stakeholder
Head of partner ecosystem / Claude enterprise leader
Initiative
Move Claude from pilots to production through clouds, SIs, and regulated-industry partners
Strategic rationale
The selling angle is not “your model is popular.” Anthropic already knows that. The higher-value angle is that partner volume, certification, and enterprise deployment quality are now operating bottlenecks. That is where a new vendor earns attention.
04 / Focused skill modes

Run the full system — or pull one part when that's all you need.

QS
Quick Scan

Fast account triage. Snapshot, top triggers, top personas, and a go / maybe / no-go call before you invest in the full run.

5 min
TR
Trigger Radar

Run the urgency layer by itself. Why-now score, ranked triggers, and next move when you need a fast read on timing.

Stage 0.5
WP
Warm Path

LinkedIn CSV flow, internal leverage, investor overlap, customer routes, and partner paths without running the full War Room Brief.

Stage 06
LI
Leadership Intel

Executive research, route selection, the Earned Right Test, and executive narrative building when the C-suite is the real question.

Stage 5.5
AR
Account Review

Audit an existing War Room Brief for slop, weak warm paths, unverified claims, and mismatched executive messaging before the field uses it.

Post-run
05 / FAQ

Short answers. No marketing voice.

What is Greenfield?
Greenfield is a free AI skill for breaking into greenfield accounts. Instead of jumping straight to an email, it makes the model research the account, test the angles, map the warm paths, and build the best way in before outreach starts.
How should I install it?
If you use ChatGPT, Gemini, or Claude in the browser, copy the full GREENFIELD.md and paste it into the chat. That file is self-contained and includes the core guardrails and voice rules. If you use Claude Code, Codex, Gemini CLI, Cursor, Antigravity, or Replit, use npx skills add mangosteen-studio/greenfield. If you want the packaged local install, use npm i -g @mangosteenstudio/greenfieldaccount and then run greenfieldaccount init.
Who is it for?
It is for AEs, SDRs, founders, and GTM teams working cold or greenfield accounts. If the problem is figuring out who to go after, what the wedge is, and how to get a reply, this is built for that.
How does it work?
You give it one target account. Then it runs the full process across triggers, internal leverage, LinkedIn paths, leadership, executive routing, technical context, and messaging until it finds the best route in. On a full run, it explicitly offers the LinkedIn Connections Co-Pilot after Trigger Radar before it branches into startup, YC, or enterprise context. The point is to leave with a real plan, not generic outreach.
Does it skip LinkedIn if I'm at a startup or YC company?
No. On a full run, Greenfield should always offer the LinkedIn Connections Co-Pilot after Trigger Radar. You can say yes, no, or later, but it should not silently skip that step just because you're at a startup, YC-backed, or founder-led company.
What do I get from one run?
You get a War Room Brief with the why-now signals, ranked warm paths, primary stakeholder, executive route, warm intro brief, first-message drafts, and a 30-day execution plan.
What is the Earned Right Test?
It is the gate that stops bad executive outreach. If the model cannot show one real, specific insight about the executive and the company situation, it does not write the executive email. It sends you back to research instead.
Where does Greenfield work best?
It works best in CLI tools like Claude Code, Codex, Cursor, Gemini CLI, Antigravity, and Replit, where the model can browse, call tools, or use connected data. It also works in web chat tools like ChatGPT, Gemini, and Claude, but you will need to provide more context manually.
Can I use just one part of it?
Yes. You can run the full system or just use one focused skill like Quick Scan, Trigger Radar, Warm Path, Leadership Intel, or Account Review when you need one part fast.
Does it verify the research?
Yes. It separates claims into Verified, Inferred, and Unverified, and it gives the run a verification score at the end. The goal is to make the output honest about what was actually confirmed.
Does it work as a standalone paste?
Yes. The main markdown file is self-contained, so you can paste it directly into an AI assistant. Web chat does not receive separate files like guardrails/VOICE.md, which is why the essential one-question rule, verification discipline, banned phrases, draft rules, and output rules are embedded directly in GREENFIELD.md. The installed repo and npm paths also include the expanded guardrail files and focused skills.
Is Greenfield free and open source?
Yes. It is free, open source, MIT-licensed, and built to run in your own workflow. There is no login, hosted app, or paywall on the core skill.

Stop staring at the list.
Open the account.

One install gets it running. One account at a time, you find the trigger, the warm path, the earned-right insight, and the wedge a generic cold email will never get you.