Latias94

This skill should be used when the user asks to "build a Fret app UI", "refactor an existing Fret UI", "pick a theme/tokens baseline", or "compose shadcn-style recipes". Provides a product-oriented workflow (style → recipes → UX polish → gates) to ship a…

Runs and triages Fret UI diagnostics via `fretboard-dev diag` (scripts, bundles, shareable artifacts, compare/triage, perf gates). Use when user asks to reproduce a flaky UI bug, write a diag script, capture/share a diagnostics bundle, triage/compare bundles,…

This skill should be used when the user asks to "optimize UI performance", "investigate jank", "create a perf baseline", or "attribute worst-frame hitches". Provides a perf workflow (tail vs typical, suite normalization, worst-bundle attribution) to land…

This skill should be used when the user asks to capture Tracy traces, profile a worst-frame hitch, or correlate `diag perf` bundle evidence with a timeline. Bridges Fret's perf gates (diag perf + bundles) with Tracy profiling: reproduce worst bundles, capture…

This skill should be used when the user asks to "turn a fix into a skill", "capture reusable learnings", "write/refresh a Fret skill", or "add a regression gate + evidence". Provides a workflow to convert tricky fixes into repeatable skills with tests/diag…

This skill should be used when the user asks to "find problems from a Fret user's perspective", "audit developer experience", "build real examples to expose framework friction", or "review onboarding/examples/scaffolds/API ergonomics". Provides a task-driven…

This skill should be used when the user asks to "review a Fret UI", "polish UX", "audit focus/overlays", or "check token drift and `test_id` stability". Provides a framework-aligned audit workflow (tokens, focus-visible, overlays, commands gating) with…

Align shadcn/ui v4 + Radix behavior and composition to Fret. Default visual parity targets the current `new-york-v4` registry source unless the user names another style; semantics still follow Radix/Base UI. Use when user says "align shadcn", "parity…

Merman release operator workflow. Use when preparing a new Merman or independently versioned support-crate release, updating changelog or release notes, bumping package versions, running release preflight, creating or verifying a tag release, dispatching…

Align Merman to a selected Mermaid release and its companion behavior graph. Use when changing the pinned Mermaid baseline, evaluating compatible or latest companion packages, admitting added or removed diagrams and layouts, porting upstream…

Prepare and validate a focused Merman pull request with change-scoped Rust, FFI, feature/dependency, generated-license, workflow, platform, and performance checks. Use when opening, updating, reviewing, or repairing a PR, deciding whether a CI gate is…

Build evidence-backed, user-facing Merman release reports by comparing a configurable base release tag and target revision. Use for alpha/beta/stable release analysis, changelog-ready summaries, dependency and Cargo feature comparisons, native/WASM/Web/Node…

Diagnose and improve Merman latency, throughput, memory use, input-amplification complexity, dependency closure, or artifact size with evidence matched to the claim and semantics-preserving implementation. Use when investigating a performance regression,…

Generate or edit raster images through a running local codex-helper or OpenAI-compatible relay, including text-to-image, image-to-image, multi-image edits, masks/inpainting, streaming image responses, batch generation, 2K/4K gpt-image-2 sizing, and…

Generate or reference-edit raster images through a running local codex-helper proxy using its OpenAI-compatible `/v1/images/generations` and JSON `/v1/images/edits` bridges. Use when Codex should create bitmap images via the user's local relay/provider chain,…

Diagnoses codex-helper and Codex failures from a Codex session key by correlating helper logs, helper configuration/state, and Codex session JSONL files. Use when the user provides a Codex session id/key or asks why Codex is waiting, stuck, disconnected,…

通用《魔兽世界》PvE 职业专精导师 skill。用于任意职业/专精的循环教学、天赋理解、技能与被动机制拆解、属性与装备思路、团本 Boss 处理、大秘境打法、Warcraft Logs 复盘、SimC/Raidbots 解释、高分玩家视角答疑；尤其适合用户问“这个专精怎么打”“为什么攻略这么循环”“我当前天赋该怎么理解”“日志哪里有问题”“大米/团本怎么提升”。不用于 Voidforge/R 币/宏伟宝库路线决策，那类问题交给 wow-voidforge-roll-planner。

规划《魔兽世界》12.0.5 晦暗虚空核心 / Voidforge bonus roll、宏伟宝库 6 代币和本周装备获取路线。用于“这周 R 币投哪个 M+ / Boss”“低保拿装备还是换核心”“是否继续追同一去重池”“用哪个拾取专精 roll”“Hero talents / 套装 / 制作装是否改变 roll 路线”等决策；BIS、属性、WCL / Mythicstats 只作为决策依据，不输出通用职业攻略。

Use when a user asks to upgrade Dear ImGui, cimgui, ImPlot, ImPlot3D, ImNodes, ImGuizmo, Dear ImGui Test Engine, or related bindings in this repository. Refresh submodules, regenerate pregenerated native/WASM bindings, audit safe Rust API and backend shim…

Use lspi (an MCP server) to give Codex LSP-powered semantic navigation and safe refactoring. Use this when you need symbol-level understanding (hover/definition/references/implementation/typeDefinition), or when doing a rename across a real codebase and you…