← ClaudeAtlas

hyperbuild-16-ship-gatelisted

Step 16 of the hyperbuild pipeline — THE ship gate (FINAL STEP). Spawns hb-gate-verifier to run the ship checklist mechanically: full test suite green, lint/analyzer clean, every task status done, every epic acceptance criterion checked, PRD coverage matrix complete, platform-appropriate build succeeds, every generated-skill scripts/*.sh gate passes, the feature→task→file→test TRACEABILITY CHAIN holds for every must/should feature, and app/ is a clean git repo with scaffold/wave/epic history. Produces runs/<run_tag>/gates/ship-report.md. The gate's verdict is final — failures are fixed by changing the app, never by re-interpreting checks; max 3 fix rounds, then blocked with an honest report. On pass, delivers the final user message. Invoked by the hyperbuild router via Skill(); not run directly by users.
zakariaf/hyperbuild · ★ 0 · AI & Automation · score 72
Install: claude install-skill zakariaf/hyperbuild
# Step 16 — Ship gate (FINAL STEP) You are executing step 16 (ship-gate) of the hyperbuild pipeline. Step 15's adversarial review has been applied and its tests were green; nothing runs after this step — when the gate passes you deliver the final message and the run is complete. **Goal:** a mechanical, evidence-backed verdict that the app is done — and the final message to the user. **THE GATE'S VERDICT IS FINAL.** You may not re-run individual checks and re-classify their failures as false positives, downgrade a check to "advisory", shrink the checklist between rounds, or memo away a failure. If a check fails, change the APP (or the genuinely stale artifact, with named evidence) until the gate passes. A blocked run with a true manifest beats a shipped app that lies. --- ## Inputs Read these before spawning: - `runs/<run_tag>/manifest.json` — run_tag, gear, platform, stage. If the router did not name the active run_tag, take the run whose manifest has `stage: BUILD`. - `runs/<run_tag>/idea.md` — verbatim app idea. GOSPEL. - `runs/<run_tag>/decisions/platform.md` — chosen platform + rationale (step 1). - `runs/<run_tag>/scaffold.md` — the `## Toolchain` section: the VERIFIED build / test / lint commands (step 13 recorded them; steps 14–16 read them verbatim and never re-derive). - `research/02-engineering/author/stack-guide.md` — the committed tooling decisions (step 5) — cross-check only; scaffold.md's verified commands win. - `research/product-spec.md` — feature list +