ops-gtm

これはレビュー時に保存された抜粋です。完全で最新の内容は外部ソースを確認してください。

---
name: ops-gtm
description: Go-to-market strategy planner. Generates a complete GTM plan across paid, unpaid, marketing, sales, and AI-automation channels for any project — and hands executable campaigns off to /marketing.
argument-hint: "[plan|paid|unpaid|sales|automation|launch|brief|setup]"
allowed-tools:
  - Bash
  - Read
  - Write
  - Grep
  - Glob
  - Agent
  - TeamCreate
  - SendMessage
  - AskUserQuestion
  - WebFetch
  - WebSearch
  - Skill
effort: medium
maxTurns: 40
---

# OPS ► GTM COMMAND CENTER

Strategy layer that sits on top of `/marketing`. `/marketing` runs campaigns; `/gtm` decides what to run, across paid, unpaid, sales, and AI-automation channels, and then hands the executable pieces to `/marketing` via the Skill tool.

## Runtime Context

Before executing, load available context:

1. **Preferences**: Read `${CLAUDE_PLUGIN_DATA_DIR:-$HOME/.claude/plugins/data/ops-ops-marketplace}/preferences.json`
   - `timezone` — timestamp all output correctly
   - `gtm_default_project`, `gtm_default_audience`, `gtm_brand_voice` — project-level defaults when set
   - `gtm_monthly_budget`, `gtm_stage` (`pre-launch|beta|ga|scale`) — used to tune channel recommendations

2. **Cached plans**: List `${CLAUDE_PLUGIN_DATA_DIR}/gtm/*.md` to surface recent plans. Never overwrite a prior plan file — always append a new dated file.

3. **Daemon health**: Read `${CLAUDE_PLUGIN_DATA_DIR}/daemon-health.json`. If `action_needed` is not null, surface it before running any long planning flow.

4. **Repo auto-scan (background)**: For the current working directory, in parallel and with `run_in_background: true`:
   - `git remote -v` → infer project slug and org
   - `cat README.md` (or `README.*`) → product description, ICP hints
   - Glob `package.json`, `pyproject.toml`, `Cargo.toml`, `go.mod` → tech stack
   - Glob `.planning/**/*.md` and `docs/**/*.md` → prior briefs, positioning notes
   - Resolve: project name, one-line pitch, primary tech, existing channels hinted in the repo

5. **`/marketing` credential probe (read-only)**: Do NOT re-resolve API keys in this skill. Instead, when the user asks to launch something, delegate to `/marketing` via the Skill tool — `/marketing` owns the credential resolution chain (userConfig → env vars → Doppler → Dashlane → Keychain → gcloud ADC).

## Agent Teams support

If `CLAUDE_CODE_EXPERIMENTAL_AGENT_TEAMS=1` is set, use **Agent Teams** for the `plan` sub-command so the four research agents share context and report progress in real-time:

```
TeamCreate("gtm-research-team")
Agent(team_name="gtm-research-team", name="paid-research",       prompt="Research paid acquisition channels that fit ${PROJECT_TYPE} at ${STAGE} with $${MONTHLY_BUDGET}/mo. Return a ranked list with fit signals, expected CAC, and which /marketing sub-command (if any) executes each channel.")
Agent(team_name="gtm-research-team", name="unpaid-research",     prompt="Research organic channels (SEO, content, community, PR, partnerships, referrals, lifecycle email) for ${PROJECT_TYPE} at ${STAGE}. Return ranked list with effort estimate, time-to-signal, and /marketing sub-command mapping.")
Agent(team_name="gtm-research-team", name="sales-research",      prompt="Design a sales motion for ${PROJECT_TYPE}: outbound vs inbound vs PLG vs channel. Propose a 30/60/90 plan with concrete activities and target metrics.")
Agent(team_name="gtm-research-team", name="automation-research", prompt="Propose AI-automation recipes: lead enrichment, personalized outreach, content ops, lifecycle copy, support deflection, lead scoring. For each, list trigger, model/tool stack, and where it plugs into /marketing.")
```

If the flag is NOT set, use standard fire-and-forget subagents with the same four prompts.

---

## Sub-command Routing

Route `$ARGUMENTS` to the correct section below:

| Input | Action |
|---|---|
| (empty), plan | Full GTM plan across all four avenues |
| paid | Paid acquisition deep-dive (Meta, Google, LinkedIn, TikTok, affiliates