Launch QA Gate

Pre-launch quality assurance gate (S9B). Audits 6 domains in parallel and produces a go/no-go decision backed by evidence before any site goes live or experiment launches.

Invocation

/lp-launch-qa --business <BIZ> [--domain conversion|seo|performance|legal|brand-copy|measurement|security|all]

Arguments:

• --business: Business unit code (BRIK, PLAT, PIPE, BOS)

Business resolution pre-flight: If --business is absent or the directory docs/business-os/strategy/<BIZ>/ does not exist, apply _shared/business-resolution.md before any other step.

• --domain: Optional scope filter. Default: all

Fast path examples:

/lp-launch-qa --business BRIK
/lp-launch-qa --business PIPE --domain conversion
/lp-launch-qa --business BRIK --domain seo

When to use: After /lp-do-build completes (DO); before /lp-launch (S10) or any experiment driving external traffic; as a periodic health check on live production sites.

When NOT to use: During development; for non-customer-facing internal tools; as a replacement for /lp-do-plan validation contracts.

Operating Mode

AUDIT + GATE (read-only)

Allowed: Read deployed site; run automated checks (Lighthouse, crawler, smoke tests); read config files (analytics IDs, sitemap, robots.txt, legal doc paths); inspect network requests; read source code.

Not allowed: Code changes, config changes, deployments, destructive commands, committing fixes.

Commits allowed: QA report artifact (docs/business-os/site-upgrades/<BIZ>/launch-qa-report-YYYY-MM-DD.md); startup loop checkpoint update (docs/business-os/startup-baselines/<BIZ>/loop-state.json).

Inputs

• --business (required) and --domain (optional, default all)
• Site baseline: docs/business-os/site-upgrades/<BIZ>/latest.user.md
• Platform capability baseline: docs/business-os/platform-capability/latest.user.md
• Startup loop state: docs/business-os/startup-baselines/<BIZ>/loop-state.json

If baseline/loop-state is missing: STOP — "Baseline missing. Run /lp-do-fact-find with Startup-Deliverable-Alias: website-upgrade-backlog to create site baseline before launch QA."

Workflow

1) Intake and discovery

1a) Validate arguments — confirm --business valid; confirm --domain is one of: conversion, seo, performance, legal, brand-copy, measurement, all.

1b) Locate site baseline and loop state — read latest.user.md for deployment URL, analytics config, conversion flows, legal doc paths; read loop state to confirm DO (/lp-do-build) complete. If DO (/lp-do-build) incomplete: STOP — "Build incomplete. Complete /lp-do-build (DO) tasks before running launch QA."

1c) Identify deployment target — staging URL for pre-launch; production URL for health check. Default to staging unless user specifies production.

1d) Read platform capability baseline — confirm analytics platform, SEO infrastructure, performance baseline expectations.

2) Execute domain checks

If --domain all (default): dispatch all 8 domain modules simultaneously via Task tool in a SINGLE message. Load modules/report-template.md for output format.

Protocol: _shared/subagent-dispatch-contract.md (Model A — domain subagents run read-only audit checks; orchestrator aggregates results).

Dispatch brief per domain subagent:

• Load modules/domain-<name>.md for check definitions
• Perform all checks; collect evidence
• Return: { domain: "<name>", status: "pass|fail|warn", checks: [{ id, status, evidence }] }
• touched_files: [] (audit-only — no file writes)

Await all 7 completions. If a domain subagent returns status: fail (unrecoverable error — not a failing check): quarantine per dispatch contract §5; flag domain as incomplete in report.

If --domain <X> (single domain): load only modules/domain-<X>.md; run checks without dispatching subagents.

Domain modules:

• modules/domain-conversion.md — C1–C5
• modules/domain-seo.md — S1–S6
• modules/domain-performance.md — P1–P5
• modules/domain-legal.md — L1–L6
• modules/domain-brand-copy.md — BC-04, BC-05, BC-07
• modules/domain-measurement.md — M1–M7
• modules/domain-security.md — SEC-01–SEC-08
• modules/domain-data-hardening.md — DH-01–DH-07 (dispatches /lp-do-data-audit)

3) Cross-domain synthesis

MANDATORY — do not skip.

Scan all domain verdict objects for:

• Cross-domain failure patterns (e.g., analytics not firing may indicate both a Conversion failure AND a Measurement failure)
• Conflicting signals between domains
• Shared root causes across multiple domain failures

Document synthesis findings in the report under "Cross-Domain Analysis" before emitting go/no-go.

4) Aggregate go/no-go decision

Validate that all 8 domain verdicts are received before emitting. Then:

• GO: all blocker checks pass; warnings documented for follow-up
• NO-GO: one or more blocker checks fail

Blocker vs. warning severity:

• Blocker: Any failure in Conversion or Legal; Performance P1–P3 failures; Measurement M1, M2, or M7 failures; Security domain status: fail (any of SEC-01–SEC-05, SEC-07, SEC-08 [high/critical CVE], or SEC-06 [unauthenticated access to private route]); Data Hardening domain status: fail (any CRITICAL or HIGH finding from /lp-do-data-audit)
• Warning: SEO failures; Performance P4–P5; all Brand Copy failures (GATE-BD-06b Warn); Measurement M3, M5, M6; all DV-series delayed checks; Security status: warn (SEC-06 missing rate-limiting when auth is correct; SEC-08 moderate/low CVEs only); Data Hardening MEDIUM/LOW findings only
• Existing-site gap advisory (non-blocking): M1/M2 failure where S1B was not run → recommend measurement-quality-audit-prompt.md

5) Write QA report

Use modules/report-template.md as output structure. Write to: docs/business-os/site-upgrades/<BIZ>/launch-qa-report-YYYY-MM-DD.md

Update docs/business-os/startup-baselines/<BIZ>/loop-state.json (S9B checkpoint: complete if GO, blocked if NO-GO). Commit both artifacts.

Quality Checks

A launch QA run is complete only if:

• All invocation arguments validated
• Site baseline and loop state located (or missing baseline reported as blocker)
• Deployment URL identified and accessible
• All selected domain checks executed (no checks skipped without documented reason; data hardening domain runs /lp-do-data-audit and awaits its PASS/FAIL verdict)
• Each check result includes pass/fail + evidence
• Cross-domain synthesis completed before go/no-go
• Blocker vs. warning severity correctly assigned
• Go/no-go decision matches blocker count
• QA report artifact written with all required sections
• Loop state updated (if applicable)
• QA report committed to repo