Release Pr Quality Gate

---
name: release-pr-quality-gate
description: Build an AgentWorks release or pull-request browser quality gate over saved suites and exact build identity. Use when a merge, deployment, or promotion needs an auditable QA decision.
---

# Release and PR Quality Gate

## Outcome

Create an auditable decision workflow that binds an exact change/build to required Browser QA suites, evaluates completeness and policy, and publishes `pass`, `fail`, or `needs_review` with evidence.

## When to use

Use after the required suites have passed independently and can run unattended against a uniquely identified preview or release candidate. Use scheduled monitoring for time-based deployed-health checks.

## Discovery and user direction

Inspect the current workflow, goals, metrics, configuration, capabilities, stores, reports, and triggers before proposing changes. Summarize reusable design and gaps, then ask focused questions for unresolved customer choices such as scope, success, approvals, thresholds, ownership, and budgets. Record the answers as customer direction. Installation alone does not approve workflow changes or execution. Default to one small-team workflow; split only for incompatible access or lifecycle boundaries.

## Required inputs

Resolve trigger type, repository/change/build/environment identity, required suites and groups, route selections, blocking and review rules, timeout/cancellation behavior, evidence retention, duplicate-run policy, and authorized status/report destinations.

## Plan and AgentWorks tools

Use deterministic preparation and finalization for identity, expected-suite set, completeness, and verdict derivation. Run existing suite routes rather than copying their tests. Persist `needs_review` decisions for later review and let a separate finalization action consume the durable decision. Configure authenticated API triggers through the supported Setup surface and integrations.

## Knowledge and persistence

Persist gate, suite-run, decision, evidence, and delivery receipts keyed by exact change/build identity and policy revision. Keep canonical tests in their source packages and reusable application facts in KB notes. Never treat attempted delivery as a published status.

## Validation and reporting

Require terminal results for every required suite/group, exact tested identity, valid evidence, and a deterministic policy decision. Missing, cancelled, stale, or mismatched results cannot pass. The dashboard shows the gate timeline, suite drill-down, tested identity, evidence, review state, history, and actual delivery receipt.

## Guardrails

Do not test repository HEAD when a different build is under review, reuse results from another commit, weaken policy after failures, auto-approve review states, or post external statuses without configured authorization. Preserve reruns as separate attempts.

## Read details when needed

- [Workflow design and outcomes](../../references/workflow-design-and-outcomes.md): goals, metrics, and current-versus-separate workflow decisions.
- [AgentWorks plan and tools](../references/agentworks-plan-and-tools.md): plan, execution, and persistence choices.
- [Evidence capture](../references/evidence-capture.md): gate evidence provenance.
- [Release gate guide](references/release-gate-workflow.md): trigger, aggregation, verdict, and delivery.
- [Example gate policy](examples/release-gate-policy.json): fictional policy shape.
- [Catalog metadata](playbook.json): presentation and optional recommendations.

## Completion contract

Return installed playbook and policy revisions, trigger and identity mapping, required suite/group set, customer overrides, controlled trial result, report and delivery locations/receipts, capability resolution, and unresolved blockers.