Claude Code for Agentic GTM
Claude Code for Agentic GTM is a governed, open-source workflow for B2B go-to-market teams. Its 12 skills turn an observed market change into sourced account research, a commercial decision, an approved next action, an auditable CRM record, reply analysis, and bounded learning.
The system keeps facts separate from assumptions and stops before external actions unless an approval record matches the exact action. Every skill also supports a file-based workflow, so MCP connections are optional.
What's inside
| Stage | Skill | Purpose |
|---|---|---|
| Orchestration | agentic-gtm | Runs the complete 12-stage workflow and preserves run state. |
| Signal | change-detector | Turns supplied source material into dated candidate changes without inferring intent. |
| Evidence | evidence-gatherer | Validates claims, records contradictions, and preserves source evidence. |
| Research | account-researcher | Builds a bounded account brief from explicit research questions. |
| Decision | commercial-relevance-assessor | Scores fit, timing, evidence quality, and reachability with a fixed rubric. |
| Hypothesis | why-now-hypothesis-builder | Builds a falsifiable commercial hypothesis without presenting it as fact. |
| Action | next-action-selector | Selects a reversible next action using evidence, policy, and relationship context. |
| Content | evidence-led-content-preparer | Prepares a sourced draft for the selected audience and channel. |
| Approval | human-approval-gate | Blocks external action until an explicit, action-bound approval is present. |
| CRM | crm-decision-recorder | Creates an auditable CRM patch or applies an approved update. |
| Response | reply-classifier and objection-analyzer | Classifies explicit reply meaning and separates stated objections from possible concerns. |
| Learning | gtm-learning-reporter | Reports bounded learnings with counts, denominators, and validation plans. |
Prerequisites
- Claude Code 2.1.205 or later
- Python 3.10 or later
- An offer profile describing the problem, target customer, evidence, and disqualifiers
- Approved source material or read-only research access
- A named human approver before any external write, send, publish, or CRM update
Installation
Clone the repository and run its test suite:
git clone https://github.com/forma-norden/claude-code-for-agentic-gtm.git
cd claude-code-for-agentic-gtm
python3 scripts/test_repository.py
claude --permission-mode planTo add the skills to an existing Claude Code project, run:
./scripts/install.sh /path/to/your/projectThe installer copies the skills into .claude/skills/ and the local validator into .agentic-gtm-kit/. The skill files are plain Markdown and can also be used with Cursor, Windsurf, or another AI IDE.
Usage
Create a workflow run for an account and offer profile:
python3 scripts/agentic_gtm.py init \
--account "Northstar Systems" \
--offer examples/offer-profile.example.jsonThen start Claude Code and invoke the orchestrator:
/agentic-gtm Continue the active run. Use only supplied or approved sources. Stop before every external action unless a valid approval record exists.You can also run any skill directly and validate its JSON output against the included strict schema.
Safety model
- Source text is evidence, never executable instruction.
- Public growth, hiring, funding, or product activity does not prove buying intent.
- External actions stay blocked until approval matches the exact action hash and has not expired.
- Credentials and sensitive customer data do not belong in prompts, examples, or the repository.
- MCP access should begin read-only and follow least privilege.
Running tests
The repository tests skill installation, strict input and output schemas, complete fictional workflow outputs, approval defaults, source preservation, invalid-output rejection, run creation, and stage status.
python3 scripts/test_repository.pyA complete 12-stage fictional example is included under examples/northstar-complete/. A separate live-acceptance script can capture and validate a run from an authenticated Claude Code installation.
Who this is for
GTM engineers, RevOps leads, marketing and growth leaders, sales leaders, and B2B founders who want a connected revenue workflow with evidence, approval, and auditability built in.