Cr8v Stacks
Strategy, Design & Liquid Performance
← All Playbooks PB-03 · ANTIGRAVITY WORKFLOWS
Free PDF Playbook · 6 Chapters · 12 pages

Engineering With Antigravity: Rules, Skills & Live Browser Gates

How Cr8v Stacks configures Antigravity workspaces (.agents/rules/, custom SKILL.md modules, subagents, and Playwright visual verification) to ship faster without regressions.

Instant full-chapter reader + downloadable PDF emailed directly · Zero spam · Privacy Policy

✓ Playbook Unlocked

The PB-03 playbook is yours.

Full chapters and checklists are open below, and your styled PDF copy is ready for instant download.

Inside · PB-03
6 Engineering Chapters
  1. Designing Global Rules & Project `.agents/rules/` Guardrails
  2. Packaging Complex Workflows Into Specialized `SKILL.md` Modules
  3. Splitting Live Reconnaissance vs. Implementation Using Subagents
  4. Pre-Release Verification: Local PHP Lint + Headless Chromium Renders
CR8V STACKS PB-03
ANTIGRAVITY WORKFLOWS
Antigravity Web Workflows

Workspace rules, modular skills, WordPress packaging automation, and live browser QA verification.

6 CH · 12 pages
Included Benchmarks
Specs & Decision Tables
  • .agents/rules/: Persistent coding, design & deployment guardrails applied automatically every turn
  • SKILL.md Engine: Reusable task SOPs for WordPress compliance, SEO/AEO checks & browser visual verification
  • Subagent Fanout: Isolated browser inspector & research agents keeping main engineering context razor sharp
Sound Familiar?

Where most builds lose speed, search visibility, or margin.

Written directly from technical audits we run before scoping client projects.

01 / GAP

Agents repeating the same mistakes across chats

Without persistent global and workspace rules, every new coding session forgets your ZIP exclusions, font rules, or OneDrive sync limits.

02 / GAP

Unverified CSS breaking mobile drawers

Code that compiles cleanly in PHP can still clip navigation labels, crop 16:9 screenshots, or trap focus behind overlays in a real browser.

03 / GAP

Slow manual packaging & OneDrive clutter

Forgetting Unix 0644 ZIP file permissions, split-media archives, or deleting temporary debug folders slows down deployments.

04 / GAP

Single-threaded prompting on big audits

Asking one huge session to inspect live competitor UX, lint PHP, and verify 10 routes bloats context quickly.

What’s Inside

6 Chapters · Field Playbook

Enter your email above to unlock the full two-column editorial playbook, interactive checklists, code blocks, and PDF download.

Unlock Full Playbook ↑
  1. 01

    Designing Global Rules & Project `.agents/rules/` Guardrails

    Turn hard-won team lessons into permanent executable constraints agents never forget.

    Locked
  2. 02

    Packaging Complex Workflows Into Specialized `SKILL.md` Modules

    Give agents step-by-step WordPress.org compliance checklists, AEO audit standards, and asset protocols.

    Locked
  3. 03

    Splitting Live Reconnaissance vs. Implementation Using Subagents

    Run live browser audits in dedicated browser subagents while keeping code context clean and fast.

    Locked
  4. 04

    Pre-Release Verification: Local PHP Lint + Headless Chromium Renders

    Test interactive forms, mega menus, PDF byte streams, and responsive grids locally before pushing.

    Locked
  5. 05

    Deterministic WordPress ZIP Packaging & Storage Hygiene

    Automate Unix permissions, split upload packs under server limits, and zero orphaned temp artifacts.

    Locked
  6. 06

    Self-Updating Engineering Loops When Requirements Evolve

    How every UX review or client feedback point permanently improves future shipped builds.

    Locked
PB-03 · ANTIGRAVITY WORKFLOWS

Shipping Production WordPress & Web Apps With Antigravity Agent Workflows

By Cr8v Stacks Engineering Team · Updated Nov 2026 · 6 Chapters · 12 pages
Download Styled PDF (12 pages) ↓
Chapter 01

Designing Global Rules & Project `.agents/rules/` Guardrails

Key Focus Turn hard-won team lessons into permanent executable constraints agents never forget.

Separate cross-project standards (light-studio visual aesthetics, storage/temp hygiene, PHP & WordPress security compliance) from project-specific engineering conventions (theme folder paths, customizer prefix keys, split-media ZIP builders) inside clean markdown rule files.

STEP 01
.agents/rules/

Persistent coding, design & deployment guardrails applied automatically.

STEP 02
SKILL.md Library

On-demand SOP checklists for compliance, SEO, and visual QA.

STEP 03
Browser Subagent

Isolated Chromium explorer audits live routes without cluttering code context.

STEP 04
Build & Lint Gate

php -l pass + automated split-media ZIP validation before push.

.agents/rules/quality_gate.md
1. Never edit core templates without running php -l across all modified PHP files.
2. Verify desktop (1440px / 1280px) and mobile (390px) rendered screenshots.
3. Run package_theme.py in repo root & keep dist-* out of OneDrive.
4. Purge scratch/ artifacts and confirm Drive C: maintains >25 GB free.
Chapter 02

Packaging Complex Workflows Into Specialized `SKILL.md` Modules

Key Focus Give agents step-by-step WordPress.org compliance checklists, AEO audit standards, and asset protocols.

Store repeatable domain playbooks—such as WordPress plugin security compliance, schema graph validation, or automated deployment packaging—inside named SKILL.md folders so agents load exact checklists right when needed without bloating everyday prompts.

Chapter 03

Splitting Live Reconnaissance vs. Implementation Using Subagents

Key Focus Run live browser audits in dedicated browser subagents while keeping code context clean and fast.

Delegate multi-step live DOM inspection, dropdown hover verification, and screenshot capture to isolated browser subagents so your main engineering context focuses cleanly on PHP, CSS, and build verification.

1
Buyer Question in H1 Direct search or prompt intent matched clearly in the top heading.
2
20–40 Word Entity Lead Self-contained definition sentence models can lift verbatim.
3
Follow-Up Questions as H2s Structured sub-headings matching buyer scope, cost & comparison queries.
4
Semantic Tables & Specs HTML <table>, <dl> & <ol> data blocks instead of vague marketing prose.
5
Named Author, Date & @graph Verified provenance + connected Organization & Service JSON-LD.
Chapter 04

Pre-Release Verification: Local PHP Lint + Headless Chromium Renders

Key Focus Test interactive forms, mega menus, PDF byte streams, and responsive grids locally before pushing.

Execute deterministic verification scripts after final edits: run `php -l` across every PHP file, render live HTML/PHP output into Playwright Chromium at desktop (1440px, 1280px) and mobile (390px) viewports, inspect real PNG captures, and validate generated REST/PDF payloads.

Chapter 05

Deterministic WordPress ZIP Packaging & Storage Hygiene

Key Focus Automate Unix permissions, split upload packs under server limits, and zero orphaned temp artifacts.

Run `package_theme.py` cleanly inside the canonical workspace repository to produce normalized core theme & media ZIPs, verify non-zero archive sizes, remove ephemeral scratch files immediately, and keep cloud sync mirrors crisp and lightweight.

Chapter 06

Self-Updating Engineering Loops When Requirements Evolve

Key Focus How every UX review or client feedback point permanently improves future shipped builds.

Whenever design feedback flags a subtle gap—such as un-animated hero cards, non-matching fonts, or missing mobile toolkits—fix the root template and enshrine the standard inside project rules immediately so future pages launch right the first time.

STEP 01

Prompt Audit

Run your top 20 buyer prompts across ChatGPT, Perplexity, Claude & Google AI.

STEP 02

Log Citations

Record whether your domain is named, which URLs are cited, and who ranks instead.

STEP 03

Compare Delta

Track month-over-month citation share and newly indexed comparison or service pages.

STEP 04

Verify UTM Traffic

Check analytics for referrals tagged utm_source=chatgpt.com and perplexity.ai.

STEP 05

30-Point AEO Scan

Run modified templates through the Cr8v Stacks Free AEO Checker after deploys.

STEP 06

Ship Next Fix

Prioritize 0-point blockers first: crawler blocks, SSR gaps, and missing schema edges.

Verified Citation & Performance Readiness (Cr8v Stacks Benchmarks)
Measured across 30 deterministic checks on cr8vstacks.com & shipped client builds (Nov 2026).
ChatGPT Search & OAI-SearchBot
94 / 100Default: 42 avg
Perplexity Live Answer Citations
92 / 100Default: 38 avg
Claude & /llms.txt Entity Parsing
96 / 100Default: 35 avg
Google AI Overviews & Core Web Vitals
95 / 100Default: 51 avg
Interactive Self-Audit

Playbook Execution Checklist

0 / 4 Complete
By The Numbers

Benchmarks from real shipped systems.

.agents/rules/
Persistent coding, design & deployment guardrails applied automatically every turn
SKILL.md Engine
Reusable task SOPs for WordPress compliance, SEO/AEO checks & browser visual verification
Subagent Fanout
Isolated browser inspector & research agents keeping main engineering context razor sharp
100% QA Gate
php -l syntax pass + real Playwright Chromium desktop/mobile screenshot verification
Architecture Layer Typical Default Approach Cr8v Stacks Playbook Standard
Ad-Hoc Chat Coding No persistent rules · Manual testing · Fragile ZIPs Regressions slip onto production servers
Antigravity System .agents/rules + SKILL.md + Playwright visual gate Consistent visual parity, clean ZIPs & verified diffs
Related Engineering Articles
Next Step

Want our team to audit your stack directly?

Bring your URL or project scope to a focused 30-minute discovery call — leave knowing exact bottlenecks, architecture options, timeline, and budget.

Book a 30-min Scope Call → All Playbooks
CR8V
ASK AI ABOUT US

Do your homework your own way

Pick a question, pick an assistant. It opens ready to go.

Ready to build what your business actually runs on?

Start Your Discovery Call
01 Pick a question
02 Open it in
© 2026 CR8V STACKS. All rights reserved.
Back to Top