team-polish

Orchestrates performance, art, audio, and QA subagents to polish a game feature for release.

Updated Aug 2, 2026
One-click install
npx skills add https://github.com/claretagrapelike32/codex-ai-game-studio --skill team-polish-claretagrapelike32
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: team-polish
Source: https://github.com/claretagrapelike32/codex-ai-game-studio/tree/main/plugins/ai-game-studio/skills/team-polish
Command: npx skills add https://github.com/claretagrapelike32/codex-ai-game-studio --skill team-polish-claretagrapelike32

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve? Polishing a game feature for release requires coordinating profiling, optimization, VFX, audio, and QA work across multiple disciplines, which is slow and error-prone when done ad hoc. This Skill runs that entire polish pipeline as a structured, gated workflow. ## Core Features & Use Cases - Multi-Agent Polish Pipeline: Delegates work to performance-analyst, engine-programmer, technical-artist, sound-designer, tools-programmer, and qa-tester subagents across six phases from assessment to sign-off. - Configurable Review Modes: Supports full, lean, and solo modes (via --review flag or production/review-mode.txt) to control how many director gate reviews are spawned. - Error Recovery Protocol: Surfaces blocked agents immediately, offers skip/retry/stop options, and always produces a partial report. - Use Case: Before shipping a new combat system, run the polish team to profile frame budget violations, optimize hotspots, polish VFX and audio mix, soak test edge cases, and receive a READY FOR RELEASE or NEEDS MORE WORK verdict. ## Quick Start Run $ai-game-studio:team-polish combat to polish the combat feature and get a release readiness report.

Frequently Asked Questions about team-polish

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I polish a game feature before release with AI agents?

Invoke the skill with the feature name, such as `$ai-game-studio:team-polish combat`. It runs a six-phase pipeline covering performance profiling, optimization, visual polish, audio polish, hardening tests, and a final release readiness sign-off.

What agents does the team-polish workflow coordinate?

It coordinates performance-analyst, engine-programmer, technical-artist, sound-designer, tools-programmer, and qa-tester subagents. Visual and audio polish phases run in parallel with optimization where the pipeline allows.

How do I change the review mode in team-polish?

Pass `--review full`, `--review lean`, or `--review solo` as an argument, or write the mode to `production/review-mode.txt`. Lean mode skips director gates except phase gates, and solo mode skips all director gate spawning.

What happens if a polish subagent gets blocked or errors?

The skill surfaces the blocked agent immediately, checks whether later phases depend on its output, and offers options to skip, retry with narrower scope, or stop. It always produces a partial report rather than discarding completed work.

Can team-polish run without user approval between phases?

No. At each phase transition it presents the subagent's proposals through the user-input mechanism, and the user must approve before the pipeline moves to the next phase.