Skills / pick-gtm-stack

/pick-gtm-stack

Most GTM stacks are assembled by vibes — a tool per blog post, overlap everywhere, gaps where it counts.

pick-gtm-stack recommends tools from a mapped, living dataset — every tool tagged by job category, customer-lifecycle stage, and real usage across 48 mapped workflows, plus the BTF core agent-infrastructure stack — so the stack for a play is recalled, not guessed.

MIT Claude Code skill tool selection lifecycle-aware

What it does

Describe a GTM task — a cold outbound play, a visitor de-anon motion, a nurture sequence, a client stack audit. The skill classifies it by lifecycle stage (awareness → education → selection → mutual commit → onboarding → retention → expansion) and the jobs it needs, then fetches the live dataset at tools.json and filters to the tools that actually serve those stages and jobs.

The output is a minimal stack: one tool per job, ranked by how many of the 48 mapped workflows already use it, each with its role and GTM impact — plus the close alternatives NOT chosen and why, so the stack doesn't bloat. Jobs no mapped tool covers get flagged honestly instead of papered over.

How it works

  1. Classify the play — which lifecycle stages it targets and which job categories it needs (sourcing, enrichment, outreach, signals, meetings…).
  2. Fetch the dataset live. curl -s https://labs.bowtiefunnel.com/tools/tools.json — it evolves with the stack; a recommendation from memory is stale by definition.
  3. Filter, then rank — lifecycle ∩ stages, category = job, ranked by workflows_used (battle-tested beats shiny).
  4. Propose the minimal stack — one tool per job, each with its role and GTM impact, plus the alternatives NOT chosen and why.
  5. Flag gaps honestly — "not in the mapped stack" + a research pointer, never an invented entry.

The dataset it recalls

browse

The GTM tool stack, mapped

The full stack with descriptions, categories, lifecycle-stage chips, and usage counts — the human-readable view of the same data.

Open the table → endpoint

tools.json

The raw dataset — fetch it from Claude Code, Cursor, or any script. Category, lifecycle, usage count, and impact per tool.

Fetch the JSON →

Red flags it kills

Install & use

# copy the skill into your Claude Code skills directory
cp -r skills/pick-gtm-stack ~/.claude/skills/

Then invoke it directly, or just describe the task:

/pick-gtm-stack

# or, in plain language —
# "what tools should I use for a cold outbound play?"
# "build a stack for de-anonymizing site visitors"
# "audit this client's tool stack for overlap"