logix

Use when planning Logix API shape from scenarios or capability atoms, managing proposal portfolios, running capability projection reviews, coordinating sub-agent planning packets, promoting repeated collisions into principles, or preparing implementation-ready plans from Logix capability SSoT documents.

在 logix 仓库规划、实施、评审或收敛任何可能影响运行时性能的需求时使用。只要用户提到性能、CI artifacts、knob snapshot、trend analyze、before/after、soak、default、231-235、pressure knob、性能旋钮、counter、counter census、kernel/runtime hot path、性能证据、性能回归、下一步优化、是否需要改 API 换性能，必须优先使用本 skill。本 skill负责指导 agent 把需求拆成小步可观测实施，获取和解读 CI 性能产物，把压力旋钮映射到内核 owner、运行时配置/逻辑路径和 counters，再决定下一步是补证据、修采集、看同分支趋势、跑显式 convergence、优化内核、还是调整 API/设计。

在 intent-flow 仓库做架构/规划/Runtime/示例/日常开发时，用最短导航定位新 docs 根骨架、archive 历史基线、代码落点与质量门

在采用 `.specify/` + `specs/` 的 Spec-Driven Development（Spec Kit）仓库中，作为“统一路由器”执行 `$speckit <stage>`，把对话里的关键裁决回写到 spec-kit 产物（`spec.md`、`plan.md`、`tasks.md`、`checklists/*`、`.specify/memory/constitution.md`）；仅在用户明确改变需求/架构/原则/澄清时触发，纯闲聊/头脑风暴保持沉默。

Use when generating, reviewing, or refactoring Logix code and you need self-contained Agent-first API rules, owner-lane decisions, React host usage, Form domain boundaries, verification control-plane boundaries, and runtime-core constraints.

Use when an Agent needs to run Logix verification from shell or CI, consume DVTools or live-derived evidence, validate a Program entry, compare before/after reports, use logix live runtime debugging tasks, or test the Agent First CLI loop.

This skill should be used when managing a tiered L1–L9 draft system rooted at docs/specs/drafts in a repository, including placing new drafts, refining and promoting drafts between levels, consolidating related drafts, and maintaining a draft index document.

用于在 intent-flow 仓库内初始化前端项目，基于内置模板（首个为 Vite + Logix Sandbox 空壳）快速生成可运行的前端工程骨架。

This skill should be used when the user asks for a comprehensive decision-making process involving external research, comparative analysis, and strategic planning. It formalizes the "Research -> Analyze -> Decide -> Plan" loop.

This skill should be used when inspecting or debugging TypeScript types in a local project, especially to understand the inferred A/E/R of an Effect or the shape of a complex generic type in a specific file.

独立的“无损修剪式”模块拆解方法论：用于在规划或实现过程中，对既有模块/超大文件（默认阈值≥1000 LOC）做渐进式无损拆解，把单体逻辑拆成互斥子模块以降低耦合与冲突面；支持两种拆分形态：单一主体用 *.*.ts 平铺、子系统用目录组织；以思想/输出协议为主、脚本为辅，可被上层 agent 嵌入任意工作流（不依赖 plan.md 等固定产物）。

Guide for creating effective skills. This skill should be used when users want to create a new skill (or update an existing skill) that extends Codex's capabilities with specialized knowledge, workflows, or tool integrations.

Install Codex skills into $CODEX_HOME/skills from a curated list or a GitHub repo path. Use when a user asks to list installable skills, install a curated skill, or install a skill from another repo (including private repos).