KilimiaoSix
@KilimiaoSix
Public Skills
design-doc-to-ui-feishu-doc
by KilimiaoSix
"Create clean rich Feishu/Lark product design documents for design-doc-to-ui runs. Use when an approved design document must be uploaded or synced to Feishu/Lark with design thinking, source traceability, user stories, design acceptance criteria, page linkage, interaction/state matrices, component tokens, accessibility notes, diagrams, tables, callouts, grids, and generated page images. Requires reading and following lark-doc, lark-doc XML/style references, and lark-whiteboard when diagrams are needed. Blocks delivery if the output would be a plain Markdown dump, screenshot directory, shallow page-by-page catalog, delivery/audit report, React/Figma/Feishu verification report, stale/failed update, mojibake-corrupted remote document, or document that fails the structured design document quality gate. Delivery materials stay local in qa artifacts and must not be uploaded to Feishu."
frontend-design-codex
by KilimiaoSix
Create distinctive, production-grade frontend interfaces in Codex. Use when the user asks to build, redesign, restyle, modernize, or visually polish web components, pages, dashboards, apps, prototypes, games, hero sections, or existing UI, especially when they want non-generic AI aesthetics, strong typography, color, motion, layout, responsive behavior, and browser-verified implementation.
design-doc-to-ui
by KilimiaoSix
"Convert vague product ideas, rough user needs, PRDs, design docs, Feishu/Lark docs, Markdown specs, prototypes, wireframes, screenshots, and brand assets into checkpointed UI design packages with concept expansion, source-backed page inventory, style/interaction design approvals, optional low/mid-fidelity interaction concept demos, imagegen UI images, review-grade structured design docs, optional Feishu rich documents with remote-link content verification, dedicated SubAgent-built runnable React prototypes, optional Figma replicas, user-feedback revisions, and 80% visual parity gates. Use when Codex needs to discuss and expand ambiguous requirements before UI design, produce interaction/visual concepts, create design documentation, generate UI visual images, build React prototypes, audit visual parity, deliver Feishu design docs, or create Figma output. Requires direct SubAgent delegation, mandatory user approval checkpoints, and companion skills for Feishu, Figma, and visual audit stages."
photo-review-workflow
by KilimiaoSix
"Use when organizing a camera photo folder and reviewing JPGs aesthetically: separate RAW files and videos into subfolders, score original JPG photos with one fixed vision-model prompt, export CSV/XLSX/PDF review artifacts, and copy full-quality JPGs into rating folders without modifying originals."
design-doc-to-ui-visual-audit
by KilimiaoSix
"Audit visual parity and baseline replication for design-doc-to-ui React and Figma outputs. Use when comparing rendered React route screenshots or Figma frames against approved AI UI images. Scores every required page across layout, components, visual style, typography, assets, and interaction state coverage. Requires each page to reach visual_similarity_score at least 0.80 and fails hard for missing pages, screenshot-only implementations, missing primary navigation/actions, broken interactions/prototype links, wrong language, generic reinterpretations, or material visual drift from approved imagegen designs."
feishu-block-surgery
by KilimiaoSix
Use for targeted Feishu Docx editing with lark-cli v2 when the user wants an existing remote Feishu document synced to local code/docs, a section or table preserved while wording is refined, or files embedded into specific table cells. Triggers include "飞书的文档帮忙也改一下", "按照原表格格式", "附件放到表格中", "先不用改源文件", and Feishu Docx URLs where structure must be preserved.
smart-search
by KilimiaoSix
基于 opencli 命令的智能搜索路由器。当用户想要使用 OpenCLI、CLI 或 API 搜索、查询、查找或研究信息时，尤其是涉及指定网站、社交媒体、技术资料、新闻、购物、旅游、求职、金融或中文内容时，务必使用此 skill
goal-prompt-builder
by KilimiaoSix
Build high-quality /goal commands for OpenAI Codex CLI 0.128+ that are audit-friendly and resistant to false completion. Use when the user wants to write, draft, generate, improve, or review a /goal prompt, including requests like "help me write a goal", "design a goal for X", "review my goal command", "make a goal for this repo", long-running Codex tasks, persistent agent objectives, Ralph loop style work, or "keep working until done". Produces a copy-pasteable /goal command using Objective, Scope, Constraints, Done when, Stop if, and an optional explicit token budget; supports step-by-step, full-description, and hybrid workflows; detects project type from local files or repo context; reads AGENTS.md/CLAUDE.md when present; and checks audit-friendliness before output.
subagent-model-router
by KilimiaoSix
Claude Code 专用：派发子 agent（Agent 工具或 Workflow agent()）前，按任务复杂度为每个子任务选择 model 参数。Use when about to spawn subagents, fan out parallel Explore/Plan/general-purpose agents, or write Workflow scripts — pick fable/opus/sonnet/haiku per task complexity, or omit model when uncertain.
design-doc-to-ui-figma-replica
by KilimiaoSix
"Create or update Figma replicas for design-doc-to-ui runs through page-level SubAgents, with at most 6 active workers. Use when a user requests Figma output that must basically replicate approved AI UI images and React screenshots as fully editable, frontend-handoff-ready frames with at least 80 percent per-page visual similarity and globally verified prototype links. Requires Figma skills, figma-use before use_figma, optional figma-generate-design for React capture, SubAgent imagegen asset reconstruction for missing illustrations/icons, and audit evidence that major elements are editable. Blocks if frames are missing, pasted full-screen images are used as implementation, major UI elements/icons/assets are not editable or replaceable, page scores are below 0.80, prototype links or cross-frame jumps for primary flows are absent, or Figma materially diverges from approved imagegen designs."
enforce-engineering-discipline
by KilimiaoSix
OpenSpec-first disciplined software engineering workflow for Codex and Claude Code. Use when implementing features, fixing bugs, refactoring, adding modules, reviewing changes, writing tests, validating Spec/OpenSpec changes, or synchronizing engineering documentation. Prefer OpenSpec as the source of truth for proposals, specs, designs, tasks, and acceptance criteria; if OpenSpec is unavailable, recommend installing it, then use the bundled manual Spec fallback to maintain equivalent process files. Do not use for one-line edits, pure Q&A, formatting-only changes, or copy edits.
design-doc-to-ui-negative-case
by KilimiaoSix
Create contrastive DesignDocToUI negative-case packages from PRDs, design docs, Markdown specs, Feishu/Lark exports, screenshots, wireframes, and product notes. Use when Codex needs to turn requirements into UI prototype concepts and design notes that deliberately expose anti-patterns, ambiguity, missing states, weak information architecture, accessibility failures, or risky design choices before proposing a corrected direction. This skill is separate from the production design-doc-to-ui workflow; use the existing design-doc-to-ui skill instead when the user asks for full gated React/Figma delivery, visual parity audits, or formal page-by-page production output.
api-doc-writer
by KilimiaoSix
Use when writing or revising product-facing API interface documentation, especially for backend REST APIs that need a concise Chinese contract document with business rules, common API conventions, per-endpoint request/response details, combined request examples, and error code explanations. Trigger when the user asks for 接口文档, API 文档, 接口说明, Feishu/Lark API docs, or to rewrite implementation-heavy notes into a pure API contract.
autonomous-multi-agent-development
by KilimiaoSix
"Autonomous multi-agent style development workflow for complex software tasks. Use when a request involves substantial feature work, cross-stack changes, debugging plus implementation, refactoring with validation, CI/deploy fixes, architecture/spec-to-code execution, or explicit requests for multi-agent, parallel, autonomous, dev-first, minimal-interruption delivery. Compatible with Codex and Claude Code: use real subagent/Task tools when available, otherwise emulate agents as named work lanes and parallel tool calls without inventing unavailable tools."