Skip to content
Back to skills

Release Gate

ASecurity

Check deployment, staging, rollback, and health verification before release. Use for user-facing, infrastructure, CI/CD, integration, or high-risk changes.

  • 2 stars
  • 0 votes
  • 0 copies
  • 2 views
  • Added September 20, 2026
ai-agentsrailsdatabaseci/cddocumentation

Security analysis

A100/100

Pro scans all 2 files and shows the line behind each finding

Scanned September 20, 2026

npx -y skills add wakqasahmed/ai-engineering-workflow-skills --skill release-gate --agent claude-code

Installs into .claude/skills of the current project.

Are you the author of Release Gate?

Add the live security badge to your README. It updates with every re-scan.

Security grade badge for Release Gate
[![Security: A — Skills Directory](https://www.skillsdirectory.com/api/skills/wakqasahmed-release-gate/badge)](https://www.skillsdirectory.com/skills/wakqasahmed-release-gate)

More formats (shields.io, HTML) on the badges page. Keep it an A: scan every change in CI with Pro.

Download with Pro
SKILL.md
---
name: release-gate
description: Check deployment, staging, rollback, and health verification before release. Use for user-facing, infrastructure, CI/CD, integration, or high-risk changes.
---

# Release Gate

Use this before staging or production release.

## Workflow

1. Confirm the reviewed commit or artifact being released.
2. Confirm environment and deployment mechanism.
3. If the release introduces or changes a required config/secret name, reconcile required config/secret names against the target environment and block on any gap (`reconcile_config`) — see [configuration preflight](references/configuration-preflight.md).
4. If a secret-delivery path is new or changed, run a non-secret canary through that path and block until the canary confirms delivery works (`run_canary`) — see [configuration preflight](references/configuration-preflight.md).
5. Run the smallest deployment validation available.
6. Smoke-test the critical route or workflow.
7. Record rollback path.
8. Create a HITL issue if deployment needs missing human-held access. This is for access that does not exist yet and only a human can create or grant it (e.g., nobody has been given the third-party account/credential itself) — distinct from step 3's reconciliation, which is for config/secret names already obtainable through an existing provisioning process that simply haven't been set in the target environment yet.

## Scope & Exemption Rules

Apply this gate selectively based on risk level to prevent release gridlock:
- **Exempt (Skip this gate)**: Low-risk changes per `AI_ENGINEERING_WORKFLOW.md` (documentation-only edits, comments, minor styling/CSS tweaks, test-only additions, and narrow bug fixes with no business-rule change, data schema, auth, or infrastructure changes) skip the full staging/rollback release gate.
- **Mandatory (Do not skip)**: Medium and High-risk changes (user-facing features, business logic changes, migrations, authentication/permissions, deployment pipelines, and third-party integrations) must satisfy all checks below before release.

## Required For Medium/High Risk

- CI status
- Review status
- Target environment
- Smoke test command
- Rollback command or previous artifact
- Health check signal
- Configuration completeness and secret-delivery verification, for changes that introduce or change required config/secrets (see [configuration preflight](references/configuration-preflight.md))

## Guardrails

- Prefer deploying reviewed artifacts over rebuilding unreviewed source.
- Do not deploy production from unreviewed PRs.
- Do not silently work around missing secrets, DNS, or account permissions.
- Test database safety (disposable/dedicated storage, staging backup timing) follows `system-level/core.md` (Test Database Safety) — this gate does not relax it.

Files in this skill

  • SKILL.md2.7 KB
  • references/configuration-preflight.md3.7 KB

Attribution

Is this your skill, or is something wrong with this listing? Request removal or report an issue. Author removals are honored within 72 hours.

Comments

Loading comments…

Newsletter

Get the weekly list of safe new skills, and the ones we pulled.