anim8gen

Generate short, simple sprite animations from natural-language prompts by using imagegen2 for raster frame candidates, agentic frame review for pose and continuity, and local anim8gen tools for alignment, validation, contact sheets, previews, and package reports. Requires the imagegen2 skill.

Generate or edit game-oriented raster assets using OpenAI gpt-image-2 through a bundled zero-dependency CLI. Use for sprites, tiles, icons, portraits, UI assets, backgrounds, reference-image edits, and iterative visual variants when the output should be a bitmap asset saved into the project.

Add a new block type to a block-diagram style app with runtime behavior, UI/editor integration, serialization, tests, docs, examples, and verification.

Create a complete example model for one or more block types, including model layout, registration, tests, screenshots, docs, and verification.

Generate a project status briefing from git state, plan reports, open checkboxes, recent commits, and worktrees when the user asks for status, current work, reports, or cleanup readiness.

Safely commit only the relevant current changes, protecting unrelated user or agent work; also handle requested push, land, or PR follow-through.

Execute a bounded ad-hoc engineering task such as a small fix, refactor, docs update, or example without the full run-plan process.

Audit and update documentation for recent changes, examples, block libraries, guides, READMEs, presentations, or newsletter entries.

Draft an executable plan file through codebase research, review, refinement, acceptance criteria, and verification planning before implementation.

Triage and fix a batch of issues from GitHub or tracker files, with reproduction, prioritization, verification, reporting, and tracker updates.

Interactively review and finalize fix-issues sprint results, including manual sign-offs, landing decisions, issue closure, tracker updates, and cleanup.

Debug one complex bug with a disciplined reproduce, trace, root-cause, fix, and verify workflow; use when guessing would be risky.

Manually verify browser or UI behavior with playwright-cli or equivalent real interactions, screenshots, snapshots, traces, and reproducible steps.

Apply readable block-diagram and state-chart layout rules such as left-to- right flow, grid alignment, spacing, orthogonal routing, and connected ports.

Build, inspect, or refresh a dashboard of plan files, classify plan status, identify next ready work, or execute a bounded number of plans.

Use playwright-cli for browser automation, UI testing, form filling, screenshots, snapshots, tracing, storage state, request mocking, and test generation.

Audit recent commits or stress-test a feature area for quality gaps, missing tests, regressions, and actionable bugs.

Refine the remaining phases of a partially executed plan against completed work, current code, drift, and updated verification needs.

Research, decompose, plan, and execute a broad goal end-to-end when the user explicitly wants autonomous planning plus implementation.

Research a broad goal, decompose it into executable sub-plans, and write a meta-plan without immediately implementing it.

Triage exported product feedback JSON, check duplicates and actionability, present recommendations, and file approved GitHub issues.

Execute one phase or all remaining phases of a plan with scoped implementation, optional worktree isolation, verification, reporting, and landing.

Maintain this Codex port of Z Skills by checking upstream changes, refreshing references/support assets, and preserving Codex-native wrappers.

Verify recent changes from actual diffs with coverage review, tests, manual UI checks when needed, fixes when requested, and residual-risk reporting.

Generate images using OpenAI's gpt-image-1 model. Ideal for game assets (sprites, tiles, UI elements, icons), concept art, and placeholder graphics. Supports iteration — "try again", "make it bluer", "go back to v1".

Generate or edit images via Google's Nano Banana (Gemini image models). Text-to-image, multi-image edit with natural-language region guidance, multi-turn iterative editing via thoughtSignature continuation. 72 built-in style presets across 10 categories. Requires GEMINI_API_KEY.