Launch
25 Claude Code subagents + 14 lifecycle skills + CLAUDE.md templates - build, ship, and monetize products with AI. Clone, run sync, done.
npx -y skills add T4LEL/Claude-Arsenal --skill launchAssembled from the repository path, not quoted from the project. Check it against their README if it does not work.
One thing to look at
- 1 stars1 stars. Stars are a popularity signal and not a quality one, but at this level it is likely that nobody has read this closely except its author, and you would be relying on your own review.
What its author says it does
Copied from the file, not written here
Use when taking a shipped product to the public - or when the user says launch, announce, Product Hunt, go public, get first users. Deploying is /ship-it; this is getting humans to show up.
SKILL.md
3.3 KB, as published. Nobody here has run it
Launch
Get humans to show up — everything gets drafted, the user posts it; never auto-publish anywhere.
Copy this checklist and check off items as you complete them:
Launch Progress:
- [ ] Step 1: Gate — production live, signup verified
- [ ] Step 2: Assets — copy, legal, analytics, OG tags
- [ ] Step 3: Channel plan
- [ ] Step 4: Posts drafted
- [ ] Step 5: Day-of runbook
- [ ] Report
Step 1 — Gate (paste real output)
Request the production URL and confirm it loads with real content. Run signup end to end and confirm it actually completes — not "looks right," an observed result. If it's not live or signup is broken, stop here and run /ship-it first; launching an unshipped product wastes the channel goodwill.
Step 2 — Assets
- Delegate to the copywriter agent with the product and audience: a landing copy pass where the headline states the customer outcome and there is exactly one CTA.
- Delegate to the legal-advisor agent with what tracking is actually implemented: privacy policy and any other legal pages, written to match the real tracking in place — not boilerplate claiming data collection that doesn't exist.
- Delegate to the data-analyst agent with the funnel: minimal events for visit, signup, activation, instrumented and confirmed firing (real event in a real dashboard, not "should work"). Any analytics or tracking keys stay in env vars, never printed in chat, code, or docs.
- Check OG/meta tags render correctly when the link is shared — paste a real preview (link debugger or equivalent), not an assumption. Fetch current docs via the context7 MCP tools for any tracking or meta-tag SDK involved.
Step 3 — Channel plan
Delegate to the growth-marketer agent with the product, audience, and Step 2 assets: channels ranked by fit and effort, the first 3 concrete actions for each, and a launch-day sequence with rough timing. Any market or conversion benchmarks the agent cites from general knowledge rather than this product's real data get labeled unverified assumptions in the report.
Step 4 — Posts
Delegate to the copywriter agent with the channel plan: draft one launch post per channel, written in that channel's native tone — never one blob cross-posted everywhere. Every draft stays a draft. The user reads it and posts it themselves; nothing here publishes on its own, and nothing gets scheduled or sent without the user's explicit go-ahead.
Step 5 — Day-of runbook
Posting order and times taken from the Step 3 sequence, what to monitor during the window (errors, signup counts — real numbers pulled from logs/dashboards, or say plainly they're not available yet), and the week-one follow-up actions (replying to comments, second-wave channels, a follow-up post).
Report
- Launch readiness: asset checklist status — what's built and verified vs. still open.
- Channel plan: ranked channels, effort, first 3 actions each.
- Posts drafted: which channels have drafts ready for the user to post.
- Metrics to watch: the funnel events and where to watch them.
- Week-one plan: the follow-up actions and rough timing.