Codex-is-all-you-need

Use only when the user explicitly invokes this skill to map unresolved product, domain, repository, or delivery unknowns before choosing a path.

Use only when the user explicitly invokes this skill to create or resume a persistent file-based goal tracker across sessions.

Use only when the user explicitly invokes this skill to stress-test an existing plan or design through one load-bearing question at a time.

Use only when the user explicitly invokes this skill to evaluate a skill with representative cases, isolated runs, independent judging, and evidence-driven edits.

Use for a concrete defect or regression that needs reproduction, root-cause diagnosis, a fix, and regression validation. Do not trigger for routine localized edits or primarily performance-focused work.

Use when the requested feature genuinely spans at least two delivery layers such as frontend, backend, API, data, scripts, or tests. Do not expand a single-layer change into a full-stack workflow.

Use to plan modernization, framework upgrades, service extraction, schema migrations, rewrites, and rollback paths.

Use when Python-specific typing, async behavior, packaging, pytest, ruff/mypy, scripting, or performance quality is central to the request. Do not trigger for generic small code edits merely written in Python.

Use when the user asks for a behavior-preserving refactor plan before implementation. Do not trigger when the user already requested a clear, scoped refactor to be implemented directly.

Use to implement a scoped multi-step change from an existing spec, handoff, issue, or explicit delivery request. Do not trigger for read-only analysis, planning, or a trivial localized edit.

Use only when the user explicitly invokes this skill to manage an approved or proposed baseline/scope change across artifacts, traceability, approvals, and handoff.

Use to create dev handoff packages and task cards from scoped SDLC materials, issues, specs, validation, and blockers.

Use to create or refine HLD from SRS, NFRs, domain boundaries, repo evidence, and architecture goals.

Use only when the user explicitly invokes this skill to plan and route a multi-artifact SDLC/ADS workstream without automatically invoking downstream skills.

Use to define measurable NFRs for performance, reliability, security, privacy, accessibility, observability, and operations.

Use to draft or refine PRDs covering product goals, scope, features, user flows, priorities, exceptions, and acceptance criteria.

仅在用户明确要求基于仓库证据生成项目功能能力地图、能力盘点或持久研究包时使用；不要用于普通代码探索、repo onboarding、Git 历史问答或只读解释。

Use only when the user explicitly invokes this skill to make a readiness-gate judgment on SDLC materials, specs, handoffs, issues, or direct-dev requests.

Use to build or audit RTM links across requirements, design, validation, tasks, PRs, commits, releases, and evidence gaps.

Use only when the user explicitly invokes this skill to establish a coordinated requirements baseline spanning business, user, and product inputs.

Use only when the user explicitly invokes this skill to decide whether work needs SDLC/ADS and which minimum materials, if any, are justified.

Use only when the user explicitly invokes this skill to coordinate a multi-artifact solution package across HLD, LLD, domain, ADR, NFR, and SPEC slices.

Use to write focused implementation SPEC slices for UI, API, data, admin, permissions, directory, observability, or release.

Use to draft or audit SRS with requirement IDs, functional requirements, acceptance criteria, constraints, NFR hooks, and traceability.

Use to create validation plans or behavior baselines linking requirements to checks, evidence, gaps, risks, and release gates.

Use when the user asks how to use Codex Next, which bundled skill fits, or whether a workflow is needed. Do not trigger for a clear task that can be completed directly.

Manual-only Office Memory V1 Lite workflow. Run when the user explicitly invokes $manage-office-memory from a project-local Skill or $office-memory:manage-office-memory from the installed plugin. Use to curate date-based short-term project memory, current AWARENESS.md, and durable MEMORY.md from explicitly selected project materials and whitelisted sources.

为 Codex 仓库设计、审计或重构 AGENTS.md 层级；用于写/改/审计 AGENTS.md，不用于 CLAUDE.md 或 Cursor 规则。

Use to design modular monolith boundaries, module layout, dependency rules, shared/platform seams, and extraction triggers.

Use to model bounded contexts, capabilities, ownership, domain data/language, allowed interactions, and forbidden dependencies.

Use to create or refine LLD for modules, interfaces, state machines, data structures, errors, transactions, and tests.

Use when converting data analysis results into a clear report with assumptions, findings, validation, limitations, and next steps.

Use when planning safe data cleaning, deduplication, missing-value handling, anomaly handling, and reproducible cleaned outputs.

Use when reviewing or creating SQL for analysis, metric definitions, joins, filters, aggregation, and validation queries.

Use when analyzing CSV, Excel, parquet, or table-like files and producing reproducible summaries.

Use to audit web/app accessibility, including semantics, keyboard flow, focus, labels, ARIA, contrast, and responsive behavior.

Use to design or review API contracts, auth, pagination, versioning, errors, compatibility, and OpenAPI/GraphQL shape.

Use to diagnose or improve build speed, cache behavior, bundle size, packaging output, and monorepo task execution.

Use to design, implement, review, or test CLI tools, flags, config precedence, terminal UX, errors, and docs.

Use for dependency upgrades, vulnerability fixes, version conflicts, lockfiles, package managers, and migration risk.