Skip to content
SkillModel-led5,758 tokens in use

release-gate

How a change ships: the full gate, an owner preview, a go/no-go with evidence, migrations before promotion, proof that the reviewed build is serving, and rollback.

680 stars on the repository MIT Licensephuryn/product-engineering · ai-shipping/skills/release-gate
claude plugin install ai-shipping@product-engineering
View source

Install command as written in the repository's README

Skill label

measured from its files

What installing release-gate costs

48tokens idle
5,758tokens in use
22,169tokens, up to

Idle is what sits in context while the skill is installed. In use is its instructions, loaded when it runs. Up to adds the files its instructions link. Counts are approximate, made offline with cl100k_base.

Model dependence Model-led

Script-ledMixedModel-led

Mostly instructions, so how well it works depends on the model that follows them.

This describes the skill's structure, how much of it is scripts. It does not say how well it works.

Touches

Nothing detected: no shell, network, file writes, git or credentials in its files.

Declared tags come from the skill's frontmatter; detected ones from its scripts, its code blocks and the commands it names in inline code. This is detected, not audited.

How it scores

Quality

How well the skill is written, counted from its own files: its frontmatter, its steps, examples, stated output and constraints, its size and whether every file it ships is one it names.How we score it →
Good
  • Frontmatter8/15
  • Structure19/25
  • Examples3/15
  • Output15/15
  • Size11/15
  • Resources15/15

Security

Starts at 100 and loses points for each thing in the skill's files that a reader should look at first.How we score it →
Excellent

No findings.

Repository health

Pre-computed score based on development activity, maintenance, community, maturity, and trend momentum.How we score it →
Good
  • Development activity100
  • Maintenance100
  • Community44
  • Maturity4
  • Momentum40

release-gate treats a release as a decision: the owner says go and the agent supplies the evidence. It defines what each shipping moment owes, from stage and phase close through deploying pushes to a production promotion, and how to prove afterwards that the reviewed commit is the one serving.

It names two branches (product and production), keeps release evidence in the product’s own records, and covers releases beyond web deploys, such as packages, extensions, installers and store listings.

What You Get

  • A stage and phase close sequence: full gate, visual checks, document-app in check mode, independent review rounds, closing push and owner preview
  • A promotion order: preconditions with evidence, go/no-go post, migration by connection string, fast-forward promotion, then verification and tag
  • Verification rules: probe something the diff adds and something it removes, request an unguessable id to prove code and schema agree, check workers and edge runtimes, smoke last
  • A release runbook for packages, extensions, installers and stores, with backend-first ordering and a leak check on the outgoing ref
  • Reference files for the checklist, go-live, the runbook and Netlify specifics

Common Use Cases

  • Pushing a reviewed phase to a branch that deploys
  • Promoting a build to production with a written go/no-go
  • Applying database migrations without letting a push race them
  • Releasing an installer or extension with its backend

What Makes It Unique

It gives each rule its principle first and the stack mechanism second, with Next.js, Supabase, Clerk and Netlify as the worked case and a relay service plus desktop installers as the second. It insists that silence is not a go, an inherited report is a claim, a waived precondition is recorded as waived, and the gate is the exit code. Static audits are proposed at each release and never run unasked.

Under The Hood

The skill is markdown with four reference files and a receipts file, and no scripts. The agent it guides runs git commands, the product’s gate and smoke scripts, hosting and database command lines such as the Railway and Supabase CLIs, and web requests to probe the live site. It may apply migrations through a production connection string after the owner’s go, push or fast-forward branches, tag releases, start a temporary development tunnel for previews, and write release records and reports. Pushes and promotions wait for the owner’s go; the skill forbids force-pushing the production branch, using real users’ credentials, or testing against production data.

Join founders building with open source

Opinionated takes, migration guides, cost-saving tips, and insights from the open source ecosystem.

Subscribe on SubstackJoin 750+ subscribers