v0.1.0MIT license
A launch checklist written for your project, not for every project.
An open-source agent skill for Claude Code and other agents that support the Agent Skills format. It reads your repository and live site, asks only what it cannot see, and writes a tracked checklist: what blocks launch, what can wait, and what does not apply to you.
Found in the project
The checklist it writes
Three example projects are shown here when JavaScript is on. The full checklist library is in the repository.
A long generic list hides the ten items that matter. Items that do not apply are ruled out, each with a reason you can read.
Where it can, the skill checks the repository and the live site first, and sets each status from what it finds.
For laws, store policies and provider terms it opens the official page when you use it, and records the address and the date.
The library holds 439 items across fifteen categories and twenty kinds of project. Pick what you are launching to see which ones stay. This picker is a simplified view; the skill also reads your project, adjusts priorities and records why each item was left out.
The picker needs JavaScript. The full library is in the repository under skills/launch-checklist/references.
An item is a task someone can finish and someone else can check. This is the shape of one item in the file the skill writes.
B1 B2http://example.com returned 200 with no redirect.An illustration of the format, using an example address.
So you can judge for yourself whether an item is worth skipping.
An item is done when the check passes, not when someone says so.
What was seen, where and when. For a law or a store rule, the official page and the day it was read.
Blocker, before launch, after launch, or not applicable with a one-line reason.
Run the skill again and it updates the same file, keeping your owners, dates and notes.
Reads the README, manifests, deployment files and live address, then asks at most five questions, each with the default it will assume.
Names the project types, the risk profile, the jurisdictions to consult, and whether a bad release can be undone.
Gathers evidence without changing anything: redirects, headers, DNS records, robots.txt, committed secrets, test coverage.
Keeps the items a project fact calls for, sets each priority and gate, and lists the rest as not applicable with a reason.
Opens the official page for every legal, store or provider fact and records the address and date. It says so when a page cannot be reached.
Saves LAUNCH_CHECKLIST.md in twelve sections, from the project overview to the go-live sign-off.
Summarises the blockers, the items with outside lead time and the assumptions to confirm, and offers a CSV export for your tracker.
Run it again later and it updates the statuses in the same file. It never starts over, because your tracking history lives there.
Some tasks wait on other organisations, so the checklist puts them first and fits the rest to your launch date.
Start what waits on others: payment provider activation, store enrolment, legal review. Decide the launch style, date and owner.
Freeze the launch scope. Run the full test pass. Restore one backup. Turn on monitoring and alerts.
Fix each finding or accept it in writing. Load test at the expected peak. Rehearse the rollback once.
Record the go or no-go decision. Confirm who is on call and who can call a rollback.
Deploy, run the smoke test, complete one real journey, watch for an hour, then announce.
Review what failed and what the checklist missed, then close the remaining items.
Rollback triggers are agreed before launch, with numbers, so nobody has to debate them under pressure.
These prompts are good starting points. Copy one, or describe your own project.
I am launching my SaaS next month. What have I missed?
Run a launch readiness check on this repository. Do not change anything.
We submit our app to the stores in two weeks. Give me only the launch blockers.
Update the launch checklist and export it to CSV for our tracker.
Use either the plugin or a copied skill folder, not both, or the skill will be listed twice.
Inside Claude Code, add the marketplace, then install the plugin.
/plugin marketplace add Achal13jain/launch-checklist-skill/plugin install launch-checklist@launch-checklistFrom a terminal, the same two steps are:
claude plugin marketplace add Achal13jain/launch-checklist-skill
claude plugin install launch-checklist@launch-checklistClaude uses the skill when a request matches it, or you can invoke it as /launch-checklist:launch-checklist.
Copy the skill into ~/.claude/skills/ for all your projects, or into a repository's .claude/skills/ to share it with a team.
git clone https://github.com/Achal13jain/launch-checklist-skill.git
mkdir -p ~/.claude/skills
cp -r launch-checklist/skills/launch-checklist ~/.claude/skills/Start a new session, then invoke it as /launch-checklist or just ask.
One command installs the skill into the agents you choose. It needs Node.js.
npx skills add Achal13jain/launch-checklist-skillChoose an agent with -a, for example -a claude-code, and add -g to install it for all your projects.
Codex reads personal skills from ~/.agents/skills and a repository's skills from .agents/skills.
git clone https://github.com/Achal13jain/launch-checklist-skill.git
mkdir -p ~/.agents/skills
cp -r launch-checklist/skills/launch-checklist ~/.agents/skills/Mention it with $launch-checklist, or run /skills to pick it.
gemini skills install https://github.com/Achal13jain/launch-checklist-skill.git --path skills/launch-checklistAdd --scope workspace to install it for the current project only, and enter /skills reload in a running session.
Copilot reads personal skills from ~/.copilot/skills and a repository's skills from .github/skills.
git clone https://github.com/Achal13jain/launch-checklist-skill.git
mkdir -p ~/.copilot/skills
cp -r launch-checklist/skills/launch-checklist ~/.copilot/skills/