[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"skill:01M72TEXRCSGS9N5H60W0036TN":3,"skill-same:01M72TEXRCSGS9N5H60W0036TN":145},{"skillNo":4,"name":5,"nameZh":6,"description":7,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":9,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":15,"categories":16,"keywords":17,"publishedAt":18,"linkOnly":19,"markdownContent":20,"markdownBody":21,"frontMatter":22,"refName":23,"commitSha":24,"skillDirectory":25,"forksCount":26,"repositoryHtmlUrl":27,"repositoryHomepageUrl":6,"ownerHtmlUrl":28,"ownerBlogUrl":29,"ownerDescription":6,"updatedAt":30,"install":31,"resources":44},"01M72TEXRCSGS9N5H60W0036TN","ywc-task-generator",null,"(ywc) Use when converting a specification into implementation tasks. Triggers: \"task 생성\", \"タスク生成\", \"spec to tasks\", \"task breakdown\", \"작업 분해\", \"仕様からタスク生成\", \"任务拆分\", \"从 spec 生成 task\", \"generar tareas\", \"desglosar spec\", \"implementation tasks\", \"스펙 분해\", or any spec-to-task decomposition request. Do not use for direct code implementation, spec review (use ywc-spec-validate), or planning without a written specification.","yongwoon\u002Fywc-agent-toolkit","yongwoon","https:\u002F\u002Favatars.githubusercontent.com\u002Fu\u002F9317029?v=4","01M72SZSF6VKANKG5QTMV4V95S","01M72SZSG8HHC3W2CV5P7EJ5AH",8,"MIT","https:\u002F\u002Fgithub.com\u002Fyongwoon\u002Fywc-agent-toolkit\u002Fblob\u002Fmain\u002Fcodex\u002Fskills\u002Fywc-task-generator\u002FSKILL.md",[],[],"2026-11-10 22:39:42",false,"---\nname: ywc-task-generator\ndescription: >-\n  (ywc) Use when converting a specification into implementation tasks. Triggers: \"task 생성\", \"タスク生成\", \"spec to tasks\", \"task breakdown\", \"작업 분해\", \"仕様からタスク生成\", \"任务拆分\", \"从 spec 生成 task\", \"generar tareas\", \"desglosar spec\", \"implementation tasks\", \"스펙 분해\", or any spec-to-task decomposition request. Do not use for direct code implementation, spec review (use ywc-spec-validate), or planning without a written specification.\n---\n\n# Task Generator\n\n**Announce at start:** \"I'm using the ywc-task-generator skill to decompose the specification into dependency-safe tasks.\"\n\n> **Summary**: Analyzes a specification and generates dependency-safe, reviewable implementation tasks. Each task is output as a structured directory (`README.md`, `task.md`, `test.md`), along with a top-level dependency graph.\n\nYou are a senior tech lead responsible for converting a specification into implementation tasks for a production web application. Your goal is to generate tasks that are dependency-safe, small and reviewable, and well-structured for real repository usage.\n\nThis skill assumes the upstream ambiguity has already been reduced to a written spec. If the user only has a discovery map or unresolved multi-session exploration notes, route them to `ywc-wayfinder` or back to `ywc-spec-ready` before decomposition.\n\nWhen the upstream spec cites a persisted `ywc-tech-research --output` artifact, keep that project-relative research path in `Spec Reference` and preserve its provenance markers instead of collapsing it into an untracked summary.\n\n## Rationalization Defense\n\nWhen tempted to bend a rule, check this table first:\n\n| Excuse | Reality |\n|---|---|\n| \"DB migration is small, I'll bundle it with the API task\" | DB migration must be its own task — in **every** mode. Safety invariant. |\n| \"User did not specify `--mode` \u002F `--granularity`, `human` is the safe default\" | Always ask. Wrong mode cascades into every task's size and bundling. |\n| \"Phase boundary is fuzzy, I'll let SEQUENCE express the order\" | Phase boundaries are **hard gates**. If only some Phase N tasks must finish before a Phase N+1 task starts, that task belongs in Phase N. |\n| \"Ownership is just a hint, the implementer will figure it out\" | Ownership is an **operational edit boundary**. Vague Ownership = parallel-execution conflicts later. |\n| \"Spec Reference is empty for this task, I'll skip the section\" | Never omit. Use `N\u002FA — housekeeping\u002Frefactor only` when there is no spec. Empty section = ambiguity. |\n| \"Library introduction is part of the feature task, no need to split\" | Library introduction is its own task — in every mode. Safety invariant. |\n| \"20+ tasks in one set is fine if they are small\" | At >20, suggest splitting the spec. A task set that does not fit in human review will not be reviewed. |\n\n**Violating the letter of these rules is violating the spirit.** Safety invariants (DB migration separation, library introduction separation, phase hard gates) have no exceptions.\n\n## Arguments\n| Argument | Default | Description |\n|---|---|---|\n| `--lang \u003Clanguage>` | _(resolved or asked)_ | Output language for task documents: `ko` \\| `ja` \\| `en` \\| `es` \\| `zh`. Backward-compatible aliases: `korean` \\| `japanese` \\| `english` \\| `spanish` \\| `chinese`. |\n| `--mode \u003Cmode>` | _(asked)_ | Canonical task granularity option: `human` \\| `llm`. |\n| `--granularity \u003Cmode>` | _(alias)_ | Backward-compatible alias for `--mode`; accept `human` \\| `llm`. If both flags are present and conflict, stop and ask the user to resolve the conflict. |\n| `--tasks-dir \u003Cpath>` | `tasks\u002F` | Root directory where task directories are written. Override to support re-plan iteration in a separate directory (e.g., `--tasks-dir tasks-v2\u002F`). |\n| `--initials \u003Cvalue>` | _(resolved or asked)_ | Collaborator namespace for new task phases. Must match `^[a-z0-9]{2,4}$`; explicit invalid values are rejected. |\n| `--spec \u003Cpath>` | required for preview flags; optional for direct generation | Project-relative path for persisted previews and approved consumption. Direct generation retains Step 1 inputs, including inline specs. Preview specs must stay under `docs\u002F` unless the project defines another safe root explicitly. |\n| `--preview-only` | off | Persist the canonical preview and its reservation identity\u002Fref. Do not write task directories or `dependency-graph.md`. |\n| `--preview-path \u003Cpath>` | `docs\u002Fywc-plans\u002F\u003Cslug>.task-preview.md` | Persisted preview artifact path. Must be repository-relative Markdown under a safe root. |\n| `--approve-preview` | off | Consume a previously approved persisted preview with matching identity. Never re-decompose on this path. |\n| `--non-interactive` | off | Required when an autonomous caller consumes a previously approved preview without interactive confirmation. |\n| `--manifest \u003Cpath>` | off | Optional repository-relative architecture contract; omitted input permits root-only discovery. |\n| `--architecture-evidence \u003Cpath>` | off | Optional normalized evidence artifact; paired with `--manifest`, never raw evidence text. |\n\n## Language Option\nWhen `--lang` is not specified, resolve task document language using the shared Codex YWC policy:\n[`..\u002Freferences\u002Flanguage-resolution.md`](..\u002Freferences\u002Flanguage-resolution.md). The order is explicit `--lang` > project `.codex\u002Fywc.json` > project guidance (`AGENTS.md`, `CODEX.md`, `CLAUDE.md`) > user `~\u002F.codex\u002Fywc.json` > ask user.\n\nThis skill supports `ko` | `ja` | `en` | `es` | `zh` and backward-compatible aliases `korean` | `japanese` | `english` | `spanish` | `chinese`. Treat `zh` \u002F `chinese` as Simplified Chinese unless the user explicitly asks for another Chinese locale. There is no skill-level output-language default; for non-English output rules, read [references\u002Flanguage-policy.md](references\u002Flanguage-policy.md).\n\n### Collaborator Initials Gate\nRead [references\u002Fcollaborator-initials.md](references\u002Fcollaborator-initials.md)\nand resolve validated initials before any scan, compaction, reservation, or\nartifact write. Use explicit value, project config, user config, then one\ninteractive Git-identity confirmation; malformed tiers are skipped. Unresolved\nnon-interactive generation returns `NEEDS_CONTEXT` without downstream work.\n\n---\n\n## Preview Approval Gate\nFor persisted previews and approved-preview writes, `--spec` must resolve inside a project-declared safe spec root; absent a project declaration, use `docs\u002F`. `--preview-only` and `--approve-preview` are mutually exclusive. The preview identity binds the reserved phase and reservation ref; see [wide-refactor-decomposition.md](references\u002Fwide-refactor-decomposition.md).\n`--preview-only` returns `DONE` with its path\u002Frevision\u002Fdigest and phase reservation for a separate caller approval step; it writes no task directory or graph.\n`--approve-preview` consumes only an explicitly approved preview whose spec,\ntasks dir, language, mode, path, revision, digest, and phase reservation all\nmatch; under the lock, verify and reuse that reservation without selecting or\ncreating another phase. It never re-decomposes or refreshes the preview. Missing\napproval, identity mismatch, or unsafe paths return `NEEDS_CONTEXT` before task writes. A matching already-approved unchanged preview consumes once without another confirmation. Newly generated previews remain unapproved. Without\neither preview flag, preserve the existing direct-generation path without\nrequiring preview approval.\n\nFor the canonical preview identity fields, wide-refactor digest invalidation\nrules (`Refactor Phase`, `Batch ID`, `Depends On`), and the consume-only\nguarantee, **read\n[references\u002Fwide-refactor-decomposition.md](references\u002Fwide-refactor-decomposition.md)**.\n\n---\n\n## Task Design Principles\n\n### 1. Reviewability\n- Each task should be small enough for the intended executor to hold the full context in one session\n- Size guideline depends on the selected **Granularity Mode** (see Step 5):\n  - `human` mode: ~15 files \u002F ~500 LOC (optimized for a single-reviewer PR)\n  - `llm` mode: ~35 files \u002F ~1,200 LOC (optimized for an LLM agent completing a vertical slice in one run)\n- See [references\u002Fgranularity-modes.md](references\u002Fgranularity-modes.md) for the full rule set\n- Treat these numeric limits as advisory reviewability guidance, never as automatic authorization to bundle work; split tasks that significantly exceed the selected mode's guideline or cross a safety, feature, ownership, or phase boundary\n- In `llm` mode, bundling is limited to one feature with exclusive Ownership and explicitly declared Shared Surfaces\n- Prefer self-contained change units that leave the codebase buildable on completion\n\n### 2. Dependency Safety\n- Tasks must be strictly ordered by dependency\n- Earlier tasks must not depend on later tasks\n- Every task must be independently implementable at its position\n- If Task B depends on Task A, Task A must fully enable Task B\n- After each task completes, the codebase must remain consistent and buildable\n\n### 3. Database Migration Separation (Safety Invariant)\n- Database migrations must always be separated into their own task — in every mode\n- Include schema\u002Fmodel definitions\n- Never mix migration and feature implementation\n- Point the migration task's `task.md` checklist at the shared schema guide so the implementer applies the eight invariants that fail deterministically when omitted: [..\u002Freferences\u002Fschema\u002Fcore.md](..\u002Freferences\u002Fschema\u002Fcore.md) plus the stack file matching the project (`prisma.md` \u002F `sql-ddl.md` \u002F `drizzle.md` \u002F `typeorm.md`)\n\n### 4. Library Introduction Separation (Safety Invariant)\n- Create a dedicated task when introducing a new library\u002Fframework — in every mode\n- Never mix library introduction and feature implementation\n\n---\n\n## Workflow\n### Step 1: Verify Input Specification\nReceive the specification from the user. It may come in one of these forms:\n\n- Markdown document (file path or inline)\n- PRD (Product Requirements Document)\n- Feature requirements description\n- Change requirements for existing code\n\nWhen `--preview-only` or `--approve-preview` is in play, the input spec must\ncome from `--spec \u003Cpath>`. Persisted preview and approved consume modes do not\naccept an implicit inline spec because the identity must be replay-safe and\npath-bound.\n\nIf the specification is unclear, ask specific questions to clarify the scope.\n\n### Step 2: Resolve Initials and Collect Project Context\n\nGather project context for task verification commands, ownership paths, and category selection.\n\n**Targets to collect:**\n- `AGENTS.md`, `CODEX.md`, `CLAUDE.md` — project rules, language policy, CI commands. Prefer Codex-native instructions (`AGENTS.md` \u002F `CODEX.md`) when they conflict with Claude-only guidance.\n- `package.json`, `Makefile`, `pyproject.toml`, etc. — actual lint\u002Ftest\u002Fbuild commands\n- Project directory structure — `src\u002F` layout, monorepo status, existing module placement\n- Resolve `--tasks-dir` as a repository-relative path; reject absolute, escaping, or symlink-escaping paths with `NEEDS_CONTEXT`.\n- Resolve the default `tasks\u002F` from the repository root. The default is independent of `--spec` location and the current subdirectory; never infer `\u003Ctasks-dir>` from the parent directory of `--spec`. Resolve the output directory once and use that same path for preview analysis, approved consumption, writes, and validation.\n- Existing resolved `\u003Ctasks-dir>` and `\u003Ctasks-dir>\u002Fdependency-graph.md` — whether tasks already exist, numbering collision check\n- Docker environment — whether commands require a `docker exec` prefix\n- `docs\u002Fubiquitous-language.md` (if it exists) — canonical domain terms and \"Synonyms to Avoid\"; task names, Implementation Steps, and Ownership paths must use canonical terms and never use synonym identifiers\n\n**When existing tasks are present:**\n- Determine the next initials-scoped PHASE by parsing `tasks\u002Fdependency-graph.md` first when it exists, then scanning **both** `tasks\u002F` and `tasks\u002Fcompleted\u002F`, plus corresponding linked-worktree sources. Completed tasks are moved out of `tasks\u002F` into `tasks\u002Fcompleted\u002F` by the executors (`ywc-sequential-executor` \u002F `ywc-parallel-executor`), so scanning `tasks\u002F` alone misses them and risks reusing a number that already exists. For resolved `yk`, only `yk-\u003Csix-digit-phase>-\u003Csequence>-...` candidates contribute to the maximum; legacy and other-initials IDs remain readable but do not advance `yk`. An empty owned scope starts at `yk-000001-010`. If the graph and directory scan disagree, continue with the union-based next owned PHASE and report the mismatch as a concern.\n- **Proactive compaction gate**: after resolving the starting number above, check the resolved `\u003Ctasks-dir>\u002Fdependency-graph.md` line count (`wc -l`; `\u003Ctasks-dir>` defaults to `tasks`). If it exceeds **300 lines**, run `python3 codex\u002Fskills\u002Fywc-task-generator\u002Fscripts\u002Fcompact-dependency-graph.py \u003Ctasks-dir>` before generating any new task. Skip compaction during `--preview-only`; defer it until the approved reservation is verified during consumption. Direct generation and approved consumption may compact before task writes. Report the before\u002Fafter line count; if it remains above 300, active\u002Fplanned work accounts for the size and no further action is needed. The gate only removes phases whose every task is in `\u003Ctasks-dir>\u002Fcompleted\u002F` and never touches a phase with outstanding or unresolvable work.\n- Enumerate `git worktree list --porcelain`; inspect the same relative path in graph, active, and completed sources, reporting inaccessible\u002Fmismatched paths.\n- Scope the union maximum to prefixed IDs matching validated initials; keep legacy IDs readable without claiming another namespace. Empty owned scope starts at `\u003Cinitials>-000001-010`, with sequence `010`.\n- For direct generation and `--preview-only`, hold one common-Git lock across scan, selection, reservation, and task\u002Fpreview writes. For `--approve-preview`, verify and reuse its digest-bound reservation under the lock; do not select or reserve another phase.\n- Identify dependency relationships with existing tasks, preferring `dependency-graph.md` when present, and reflect them in the new tasks' `Depends On`\n\n### Step 3: Spec Review\nReview the specification for completeness and verify that sufficient information exists for decomposition. The goal is not to demand a perfect spec, but to confirm that the key information needed to determine task boundaries is present.\n\n**When unclear areas are found:**\n- Ask the user specific questions before starting task decomposition\n- Frame questions as choices (\"Is it A or B?\") rather than vague observations (\"This part is unclear\")\n- If the spec is sufficiently clear, skip this step and proceed directly\n\n### Step 4: Confirm Language\n\nIf `--lang` is provided, skip this step. Otherwise, resolve language using\n[`..\u002Freferences\u002Flanguage-resolution.md`](..\u002Freferences\u002Flanguage-resolution.md):\nproject `.codex\u002Fywc.json`, then project guidance (`AGENTS.md`, `CODEX.md`,\n`CLAUDE.md`), then user `~\u002F.codex\u002Fywc.json`, then ask if unresolved.\n\n> \"Which language should the task documents be written in? (`ko`, `ja`, `en`, `es`, or `zh`)\"\n\n### Step 5: Confirm Granularity Mode\n\nIf `--mode` is provided, use it and skip this confirmation. If `--granularity` is provided instead, treat it as a backward-compatible alias for `--mode` and skip this confirmation. If both are provided with different values, stop and ask the user to choose one mode before decomposing tasks.\n\nIf neither flag is provided, **always ask** the user which granularity mode to apply. Do not silently default — the correct mode depends on who will execute the tasks.\n\n> \"Which mode should the tasks be generated in?\n> - `human` — reviewable units (~15 files \u002F ~500 LOC)\n> - `llm` — larger vertical slices optimized for a single LLM agent run (~35 files \u002F ~1,200 LOC)\"\n\n**Mode selection criteria** (share with the user if they are unsure):\n- Choose `human` when a person will implement and code-review each task in sequence\n- Choose `llm` when tasks will be executed autonomously by an LLM agent (Codex, etc.) in isolated worktrees, and reviewed in aggregate after completion\n\nSafety invariants — DB migration separation, Library introduction separation, Phase hard gate, post-task buildability, and single phase per task — apply in **both modes**. Only advisory size and safe internal bundling differ; mode selection never authorizes cross-feature or invariant bundling.\n\nSee [references\u002Fgranularity-modes.md](references\u002Fgranularity-modes.md) for the full mode specification.\n\n**Persist the decision.** After parsing `--mode` \u002F `--granularity` or receiving confirmation, record the selected mode in a re-checkable location (e.g., a top-line scratchpad note, or `write_memory(\"granularity_mode\", \u003Cvalue>)` when Serena MCP is available) and apply it consistently through Steps 6–9 (size, bundling, Ownership scope, Implementation Steps depth, test.md inclusion). Mode drift mid-generation produces inconsistent task directories.\n\n### Step 5.5: Resolve Architecture Contract Packet\n\nIf `--manifest` or `--architecture-evidence` is supplied, resolve the pair\nthrough the shared `..\u002Fscripts\u002Farchitecture-invariants.py` helper before task\ndecomposition. A supplied manifest is repository-relative; invalid or missing\ninput returns `NEEDS_CONTEXT` without discovery fallback. Evidence without a\nvalid manifest is `NEEDS_CONTEXT`. A valid manifest without evidence preserves\nthe existing task-generation flow and adds no packet. Derive the bounded\n`--changed-path` set only from paths already named by the specification and\ncandidate task Ownership; never scan the repository broadly.\n\nRecord only this sanitized packet in preview\u002Ftask metadata and checklist\ncontext:\n\n```text\nArchitecture Contract Packet:\n- contract_state: \u003CVALIDATED | N\u002FA — no architecture contract | NEEDS_CONTEXT>\n- component_ids: \u003Caffected component ids>\n- rule_ids: \u003Caffected rule ids>\n- verifier_requirement: v1 validation-only; no verifier execution\n- invariant_verdict: \u003CMAINTAINED | VIOLATED | N\u002FA | NEEDS_CONTEXT>\n- evidence_artifact_path: \u003Crepository-relative path or N\u002FA>\n```\nMapping is separate: `status` controls dispatch; `aggregate_verdict` becomes `invariant_verdict`; rule IDs, sanitized evidence paths, and artifact path remain distinct. Successful audits use `contract_state: VALIDATED`; no-manifest is `N\u002FA — no architecture contract`; `NEEDS_CONTEXT` is terminal before writing.\nThe successful helper integration must provide `component_ids` from its bounded changed-path mapping, `contract_state: VALIDATED`, and the normalized `.ywc-architecture-invariants-evidence.json` as `evidence_artifact_path`; missing fields are `NEEDS_CONTEXT`, never values inferred from raw inputs. `verifier_requirement` is a fixed constant rather than a helper output: every rendered packet must carry the literal `v1 validation-only; no verifier execution` value, and a packet missing that line is incomplete. The `## Architecture Contract Packet` sections in `references\u002FREADME.md.template` and `references\u002Ftask.md.template` are conditional blocks — render them only when a validated packet exists for that task; when no `--manifest` is supplied, or a valid manifest is supplied without evidence, omit the heading and every field instead of writing an empty or placeholder-filled section.\nDo not copy manifest\u002Fevidence contents, raw evidence, command-like fields,\ntranscripts, full diffs, or inferred edges into task artifacts. Propagate\n`NEEDS_CONTEXT` before writing any preview or task directory. Surface\n`VIOLATED` as a task-generation finding; `N\u002FA` and `MAINTAINED` permit the\nexisting decomposition. The packet does not elevate task-generator authority\nor authorize execution. For the conditional task-owned quality packet, read [references\u002Ftask-metadata-rules.md](references\u002Ftask-metadata-rules.md) and apply its exact Ownership, required IDs\u002Fdigests, bounded evidence paths, redaction, omission, and `NEEDS_CONTEXT` rules; Task Verify remains separate.\n\n### Step 6: Task Decomposition\n\nAnalyze the specification and decompose it into tasks following the Task Design Principles and the selected Granularity Mode.\n\n**Scale guidelines** (task counts scale down in `llm` mode due to vertical bundling):\n- **Small** (1–3 tasks expected): A single phase is sufficient. Do not force multiple phases.\n- **Medium** (4–15 tasks in `human` mode, ~3–8 in `llm` mode): Organize into 2–3 phases. This is the typical case.\n- **Large** (15+ tasks in `human` mode, 8+ in `llm` mode): Suggest to the user that the spec itself be split into multiple features. A single task set exceeding 20 tasks becomes difficult to manage.\n\n#### Planning Advisor (optional, Pattern C)\n\nFor **Medium** and **Large** specs, phase boundary decisions and task size splits benefit from frontier judgment. A wrong Phase boundary cascades into every subsequent dependency declaration and is expensive to undo once the task directories are generated.\n\nThis skill applies **Pattern C** from [advisor-pattern.md](..\u002Freferences\u002Fadvisor-pattern.md): a **single** upfront higher-capability advisor call before writing any task directories. The executor handles everything else — Task Naming, Directory Generation, Dependency Graph, and Final Validation.\n\n**When to invoke**:\n- Scale is Medium (4–15 tasks) or Large (15+ tasks), AND\n- At least one of: phase boundary is non-obvious, multiple DB migrations or library introductions compete for Phase 1, or the spec touches more than two concurrent feature areas.\n- **Skip for Small specs** (1–3 tasks) — a single-phase decomposition is obvious and does not benefit from frontier reasoning. Adding an advisor call here wastes budget.\n\n**How to invoke**: Use Codex subagent delegation when the current session exposes a delegation tool. Request one higher-capability advisor pass with the bounded payload below. If no delegation tool is available, run the same advisor checklist inline as a separate bounded pass and record the fallback in the final handoff. Do not use Claude Code-only `Task` fields, `subagent_type`, or explicit Claude model pins in the Codex bundle.\n\n- **Spec summary** — ≤20 lines distilled from your Spec Review in Step 3 (not the full spec).\n- **First-pass task list** — task name + one-line description for each candidate task from your Step 6 decomposition.\n- **Selected mode and matching guideline** — the selected `human` or `llm` mode and exactly one corresponding advisory: `human` → ~15 files \u002F ~500 LOC; `llm` → ~35 files \u002F ~1,200 LOC. Do not evaluate a task against the other mode's guideline.\n- **Known conflicts \u002F shared surfaces** — any Ownership overlaps or Shared Surfaces you identified.\n- **Project context essentials** — monorepo structure, existing phases in `tasks\u002F`, tech-stack constraints affecting dependency order.\n\nAsk the advisor for three things:\n\n1. **Phase boundary recommendations** — which tasks belong to which phase, and why each boundary is a hard gate.\n2. **Task size verification** — whether any task exceeds the selected mode's matching advisory guideline and, independently, whether a feature, Ownership, Shared Surface, safety-invariant, phase, or buildability boundary requires a split even when the numeric limit is met.\n3. **Dependency cycle risk** — any tasks with circular implicit dependencies the first-pass missed.\n\n**Budget**: exactly **1** advisor pass per invocation of this skill. Pattern C explicitly rules out re-invocation during execution — if the initial plan proves wrong, re-run the whole skill with refined input rather than calling another advisor mid-generation. This rule exists because mid-generation re-planning leads to inconsistent task directories; a fresh start is cleaner and more auditable.\n\n**Payload rules**: the summary and task list together must not exceed ~200 lines of payload. If the first-pass decomposition is already larger, narrow it before the advisor call (split the spec into sub-specs and plan each separately). Never forward the full spec verbatim.\n\n**Advisor output format** (≤300 words):\n- Phase boundary plan with one-line rationale per boundary\n- Task size concerns (if any)\n- Cycle\u002Fordering warnings (if any)\n- Single \"proceed\" or \"reconsider with refinements\" verdict\n\nAfter receiving the verdict, the executor either continues to Step 7 (Task Naming) with the adjusted plan, or surfaces the \"reconsider\" verdict to the user for refinement before proceeding.\n\n### Step 7: Task Naming\nEach new task name follows this format:\n\n```\n[INITIALS]-[PHASE]-[SEQUENCE]-[CATEGORY]-[SHORT-DESCRIPTION]\n```\n> Example: `yk-000001-010-db-create-user-table` = INITIALS `yk` + PHASE `000001` + SEQUENCE `010` + category + description.\n\n**Numbering Rules:**\n- PHASE: **6-digit** number (`000001`, `000002`, ...) — the wider width reserves headroom for multi-year project growth\n- INITIALS: validated **2–4 lowercase alphanumeric** characters resolved before allocation\n- SEQUENCE: 3-digit number (`010`, `020`, `030`, ...)\n- Sequence increments by 10 (allows inserting tasks later without renumbering)\n- Always use hyphen (`-`) to separate PHASE and SEQUENCE for readability\n\n**Category:**\n- `lib` — New library\u002Fframework introduction\n- `db` — Database migration, schema change\n- `api` — API endpoint implementation\n- `domain` — Business logic, service layer\n- `worker` — Background job, async processing\n- `ui` — Frontend component, page\n- `test` — Test-only task (e2e, load test, etc.)\n- `refactor` — Code structure improvement with no behavior change\n- `infra` — CI\u002FCD, deployment, configuration\n- `config` — Environment variables, feature flags, `.env` setup (distinct from infra: infra covers CI\u002FCD pipelines and deployment, config covers application-level settings)\n\n**Naming Quality Rules:**\n- Use only lowercase and hyphens\n- Keep names concise, descriptive, and filesystem-friendly\n- Express exactly one primary concern (in `llm` mode, \"one primary concern\" may cover a vertical slice of a single feature)\n- Avoid vague names\n- Do not use \"and\" in task names — use a unifying noun instead (e.g., `user-auth` rather than `registration-and-login`)\n\n**Examples:**\n- `yk-000001-010-db-create-user-table`\n- `yk-000001-020-lib-setup-auth-library`\n- `yk-000001-030-api-user-registration`\n- `yk-000002-010-ui-login-form`\n\nExisting numeric and unprefixed task IDs remain valid in dependencies, graph\nparsing, compaction, executor ranges, and completion moves. Do not rename\nlegacy directories or count them as another collaborator's prefixed namespace.\n\n### Step 8: Phase Organization\n- PHASE represents a meaningful dependency stage (e.g., `000001` = foundation, `000002` = core features, `000003` = integration)\n- Do not create phases arbitrarily — a new phase means \"all tasks in the previous phase must be complete before this phase can start\"\n- Tasks within the same phase may have internal ordering via SEQUENCE, but conceptually belong to the same stage\n\n**Phase Boundary Rules:**\n- Phase boundaries are **hard gates**. No task in Phase N+1 can start before all tasks in Phase N are complete\n- Dependencies between tasks within the same phase are allowed (order expressed via SEQUENCE)\n- If only some tasks in Phase N need to be complete before a specific Phase N+1 task can start, that task should be placed in Phase N instead\n- Criterion for splitting phases: \"Do all outputs from this stage need to be in place for the next stage's work to be meaningful?\"\n\n### Step 9: Directory and File Generation\nUse the pre-resolved `--tasks-dir` path for every task output. If omitted, use the repository-root `tasks\u002F`; never derive it from the spec's parent. With `--preview-only`, write the preview and its digest-bound reservation ref, but no task files or graph. With `--approve-preview`, create task files only after explicit approval and identity, reservation, authorization, and containment checks pass. Without preview flags, preserve direct generation after existing validation; no preview approval is required.\n\nGenerate the following structure for each task:\n\n```\ntasks\u002F[TASK_NAME]\u002F\n├── README.md\n├── task.md\n└── test.md (optional — included when manual verification is needed)\n```\n\nRefer to templates in the `references\u002F` directory when writing files:\n- `references\u002FREADME.md.template` — Task overview and dependency documentation\n- `references\u002Ftask.md.template` — Implementation checklist\n- `references\u002Ftest.md.template` — Manual test plan\n- `references\u002Fdependency-graph.md.template` — Top-level dependency summary\n\n#### README.md Core Elements\n\nEach README.md must include the following:\n\n- **Purpose**: What this task achieves\n- **Scope**: What is included\n- **Spec Reference**: Specification\u002Fdesign documents that must be consulted before implementation (see \"Spec Reference Rules\" below)\n- **Dependencies**: Bidirectional dependency tracking\n  - **Depends On**: Predecessor tasks this task depends on (specify what each provides)\n  - **Depended By**: Successor tasks that depend on this task (specify what each needs)\n- **Key Files**: List of files expected to be created or modified\n- **Notes**: Important considerations, design decisions\n- **Out of Scope**: What is intentionally excluded\n- **Parallel Execution Metadata**:\n  - **Model Hint**: Exactly one advisory tier and evidence-based rationale per task; use [task-metadata-rules.md](references\u002Ftask-metadata-rules.md) as the canonical tier definition.\n  - **Ownership**: Files, directories, modules, contracts this task may modify\n  - **Shared Surfaces**: Shared boundaries that may cause merge or design conflicts\n  - **Conflicts With**: Tasks that must not run in parallel\n  - **Parallelizable After**: Minimum merged baseline before isolated execution can safely start\n  - **Task Verify**: Exact commands that prove this task is complete\n\nBidirectional dependency tracking matters because developers executing a task can immediately see which downstream tasks are affected, allowing them to consider downstream requirements when designing interfaces.\n\n#### Spec Reference Rules\n\nEvery task's README.md must include a `## Spec Reference` section with three subsections: **Primary Sources** (links to source-of-truth docs, project-relative by default), **Summary** (2–5 sentence orientation), and **Out of Scope (from spec)** (scope-creep guardrail).\n\n**Hard rule**: never leave Spec Reference empty or omitted — use `N\u002FA — no external spec (housekeeping \u002F refactor \u002F config only)` when no spec exists.\n\nFor the full structure, the external-URL policy, the link-vs-summary drift protocol, and the cross-task scope-handoff convention, **read [references\u002Fspec-reference-rules.md](references\u002Fspec-reference-rules.md)**.\n\n#### Parallel Execution Metadata Rules\n\nRead [references\u002Ftask-metadata-rules.md](references\u002Ftask-metadata-rules.md) for advisory tier selection and its runtime boundaries. Every task requires a Model Hint, including tasks that will not run in parallel.\n\nWhen tasks may be executed via `git worktree`, Codex, or other isolated worker setups, every task's README.md must declare:\n\n- **Ownership** — concrete edit boundary (file paths, globs, modules, contracts). Bad: `backend`, `frontend related files`. Good: `api\u002Fsrc\u002Froutes\u002Fmetrics.ts`, `OpenAPI contract: \u002Fapi\u002Fv1\u002Fmetrics`.\n- **Shared Surfaces** — shared boundaries (DB schema, API contracts, event payloads, workspace config, CI config) that may cause conflicts even without direct file overlap.\n- **Conflicts With** — tasks that must not run in parallel even when dependencies allow. Write `(None identified)` if none.\n- **Parallelizable After** — the minimum merged baseline for safely starting isolated execution (often, but not always, the same as `Depends On`).\n- **Task Verify** — task-specific commands that prove completion (avoid project-wide gates only).\n\nFor the full rule set including good\u002Fbad Ownership examples, the complete shared-surfaces taxonomy, and the dependency-graph Parallel Execution Notes format, **read [references\u002Fparallel-execution-metadata.md](references\u002Fparallel-execution-metadata.md)**.\n\n#### task.md Core Elements\n\nEach task.md must include the following:\n\n- **Prerequisites**: Predecessor task completion status to verify before starting (checkbox format)\n- **Allowed Edit Scope**: A brief restatement of the `README.md` Ownership as an operational boundary\n- **Stop Conditions**: Conditions under which the implementer should stop and report rather than continuing\n- **Implementation Steps**: Specific implementation steps (checkbox format)\n  - Each step must reference specific files, modules, functions, or behaviors\n  - Do not use generic placeholders like \"implement core logic\" or \"handle edge cases\"\n  - Example: `Create src\u002Fmodels\u002Fuser.ts with User entity definition`\n- **Task Verify**: Task-specific verification command checklist\n- **Verification**: Confirm lint, typecheck, test, and build pass (use the project's actual commands)\nFor a `ui` task changing end-user layout\u002Fstyles, link to [`mobile-first-ui.md`](..\u002Freferences\u002Fmobile-first-ui.md) and order Implementation Steps narrowest-viewport base first, then outward `min-width` expansion. Require a PC\u002Ftablet-only exception only when explicitly stated; do not apply this to logic-only UI, admin\u002Finternal, non-UI, or unchanged legacy desktop-first CSS (no retrofit).\n**Duplicate-sensitive write Task Verify rule**: When a task touches payment, order creation, provisioning, credit \u002F balance \u002F stock \u002F quota mutation, or another duplicate-sensitive write flow, its `task.md` must include concrete Task Verify commands or test scenarios for concurrent request behavior, transaction rollback or equivalent consistency rollback on partial failure, and idempotency retry \u002F double-click behavior. The task must name the selected mechanism where applicable: atomic conditional update, row lock, optimistic lock, idempotency key, unique constraint, or persisted request\u002Fresult record. Expected results must be observable: lock\u002Fversion conflict response, exhausted stock\u002Fbalance\u002Fquota response, rollback state after a forced mid-flow failure, and duplicate retry returning the prior result without repeating the side effect. If the project has no practical local concurrency harness, the task may record a named exception only with replacement verification such as a code-level lock\u002Ftransaction proof or an integration-test plan.\n\n#### test.md (optional)\n\nWrite structured scenario-based tests (Steps + Expected Result).\n\n**Include when:**\n- UI component or screen behavior verification → always include\n- External service integration (S3, Email, Payment, etc.) → always include\n- Browser-dependent features → always include\n- Complex user workflows (multi-step) → recommended\n\n**May omit when:**\n- Pure API endpoints (coverable by unit + integration tests)\n- Database migrations (schema application verified via task.md Verification)\n- Library introduction tasks (import and basic operation verified via Verification)\n\n### Step 10: Generate Dependency Graph\nAfter generating tasks in direct-generation or approved-consumption mode, create `\u003Ctasks-dir>\u002Fdependency-graph.md` at the top level (where `\u003Ctasks-dir>` is the value of `--tasks-dir`, defaulting to `tasks\u002F`). This file serves as the single source of truth for execution order. Preview-only stops after writing and validating its preview artifact; it does not create or modify the dependency graph.\n\nRefer to `references\u002Fdependency-graph.md.template` for format. List tasks by phase and express each task's dependencies using arrow notation.\nThis graph must be consistent with the Dependencies sections in individual README.md files.\nOnce every task in a phase reaches `\u003Ctasks-dir>\u002Fcompleted\u002F`, the shared completion marker automatically compacts that phase (see `codex\u002Fskills\u002Fywc-task-generator\u002Fscripts\u002Fcompact-dependency-graph.py`) — no special markup is required; the script reads the `## Phase NNNNNN`, `## Parallel Execution Notes`, and `## Visual Dependency Graph` headings directly.\n### Step 11: Final Validation\nBefore returning a preview, direct task set, or consumed task set, apply [requirement coverage and output validation](references\u002Foutput-and-validation.md), including its attributable coverage summary and interface\u002Fverification mapping. A clean preview returns `DONE` with its preview identity; approval is a separate caller step. Then verify the following, applying task-artifact checks only to direct generation or approved consumption:\n\n**Dependency & Structure:**\n- [ ] No forward dependencies (earlier tasks do not depend on later tasks)\n- [ ] No circular dependencies\n- [ ] Phase boundaries correctly set as hard gates\n- [ ] Database migrations and library introductions separated into their own tasks (Safety Invariant in both modes)\n- [ ] No numbering collisions across the resolved graph, active\u002Fcompleted tasks, and every linked worktree source\n- [ ] Graph\u002Fdirectory numbering mismatches reported as concerns (if any)\n**Naming & Size:**\n- [ ] Naming convention followed (`[INITIALS:2-4]-[PHASE:6]-[SEQUENCE:3]-[CATEGORY]-[SHORT-DESCRIPTION]`)\n- [ ] Direct\u002Fpreview generation reserves under the common-Git lock; approved consumption verifies and reuses the preview's exact phase reservation under that lock without selecting another phase.\n- [ ] PHASE is 6 digits, SEQUENCE is 3 digits, separated by hyphen\n- [ ] No task name contains \"and\" (single concern check)\n- [ ] Each task fits within the selected Granularity Mode's size guideline\n- [ ] Selected Granularity Mode applied consistently (bundling, Implementation Steps depth, Ownership scope)\n**README.md Quality:**\n- [ ] Both \"Depends On\" and \"Depended By\" included\n- [ ] Parallel Execution Metadata included (Model Hint, Ownership, Shared Surfaces, Conflicts With, Parallelizable After, Task Verify)\n- [ ] Ownership is narrow enough to serve as an actual edit boundary, not just a restatement of Key Files\n- [ ] **Spec Reference** section included — Primary Sources, Summary, and Out of Scope (from spec) all filled in or explicitly marked `N\u002FA`\n- [ ] If Primary Sources include external URLs, only if the project allows it (default is project-relative paths only)\n\n**task.md Quality:**\n- [ ] Specific Implementation Steps included (no generic placeholders)\n- [ ] Prerequisites section included\n- [ ] Allowed Edit Scope and Stop Conditions included\n- [ ] Verification commands match the project's actual commands (based on context collected in Step 2)\n- [ ] Duplicate-sensitive write tasks include concrete concurrent request, rollback, and idempotency retry verification or a named no-harness exception with replacement evidence\n\n**Consistency:**\n- [ ] Every task has a valid advisory Model Hint and rationale; preview, README, and final summary agree before writes.\n- [ ] Dependency Graph Summary matches individual README.md files\n- [ ] If the task set is intended for parallel execution, Dependency Graph includes Parallel Execution Notes\n- [ ] Codebase is buildable after each task completes\n- [ ] Selected language's writing rules are followed\n\n---\n\n## Output Format\n\nThe final output includes:\n\n1. **Task list summary** organized by phase with Task ID, Title, and Model Hint (tier plus rationale) for every task, matching the preview when used and every generated README\n2. **Directory generation** under `\u003Ctasks-dir>\u002F`\n3. **Dependency Graph** at `\u003Ctasks-dir>\u002Fdependency-graph.md`, including Parallel Execution Notes when needed\n\nPreview-only returns the task summary and persisted preview identity; it does not claim that task directories or a dependency graph were generated. Items 2 and 3 apply only to direct generation or approved consumption.\n\nWhen parallel execution is expected, verify that each task is safe for isolated worktree or agent execution.\n\n---\n\n## Validation\n\nFor allocation, run:\n```bash\nRESOLVER_LAUNCHER=\"$(bash \"${CODEX_HOME:-$HOME\u002F.codex}\u002Fskills\u002Fscripts\u002Fselect-resolver-launcher.sh\" 2>&1)\" || { echo \"BLOCKED: ${RESOLVER_LAUNCHER#BLOCKED: }\" >&2; exit 3; }\nINITIALS_TEST=\"$(bash \"$RESOLVER_LAUNCHER\" \"ywc-task-generator\u002Fscripts\u002Ftest-initials-allocation.sh\" bash)\" || exit $?\nPARSER_TEST=\"$(bash \"$RESOLVER_LAUNCHER\" \"ywc-task-generator\u002Fscripts\u002Ftest-parser-ids.sh\" bash)\" || exit $?\nbash \"$INITIALS_TEST\" && bash \"$PARSER_TEST\"\n```\nTask Verify must cover config precedence\u002Fmalformed tiers, linked-worktree\ngraph\u002Factive\u002Fcompleted sources, scoped empty-graph maxima, concurrent distinct\nreservations, and missing-initials `NEEDS_CONTEXT` before any write.\n\nBefore returning `DONE`, verify:\n\n- [ ] The spec or plan source was read and each task traces to a source requirement.\n- [ ] For preview-only, verify the persisted preview artifact and its identity, and verify that no task directory or dependency graph was created or modified.\n- [ ] For direct generation or approved consumption, verify every emitted task directory contains `README.md` and `task.md`, and `\u003Ctasks-dir>\u002Fdependency-graph.md` exists. All task artifacts must be under the pre-resolved `\u003Ctasks-dir>`; never beside the spec unless explicitly selected.\n- [ ] UI or behavior-changing tasks have a `test.md` or an explicit test exception.\n- [ ] For direct generation or approved consumption, `dependency-graph.md` exists and matches the generated task directories.\n- [ ] Parallel Execution Notes are present when tasks are intended for parallel execution.\n- [ ] The selected output language matches `--lang` or the shared YWC language resolution result.\n- [ ] The selected Granularity Mode matches `--mode`, `--granularity`, or the user's explicit confirmation.\n- [ ] The final response reports the generated directory, task count, graph path, and `Status: DONE | DONE_WITH_CONCERNS | BLOCKED | NEEDS_CONTEXT`.\n\n## Task Execution Convention\n\nFor execution rules after task generation (completion handling, directory moves, etc.), see `references\u002Fexecution-convention.md`.\n","# Task Generator\n\n**Announce at start:** \"I'm using the ywc-task-generator skill to decompose the specification into dependency-safe tasks.\"\n\n> **Summary**: Analyzes a specification and generates dependency-safe, reviewable implementation tasks. Each task is output as a structured directory (`README.md`, `task.md`, `test.md`), along with a top-level dependency graph.\n\nYou are a senior tech lead responsible for converting a specification into implementation tasks for a production web application. Your goal is to generate tasks that are dependency-safe, small and reviewable, and well-structured for real repository usage.\n\nThis skill assumes the upstream ambiguity has already been reduced to a written spec. If the user only has a discovery map or unresolved multi-session exploration notes, route them to `ywc-wayfinder` or back to `ywc-spec-ready` before decomposition.\n\nWhen the upstream spec cites a persisted `ywc-tech-research --output` artifact, keep that project-relative research path in `Spec Reference` and preserve its provenance markers instead of collapsing it into an untracked summary.\n\n## Rationalization Defense\n\nWhen tempted to bend a rule, check this table first:\n\n| Excuse | Reality |\n|---|---|\n| \"DB migration is small, I'll bundle it with the API task\" | DB migration must be its own task — in **every** mode. Safety invariant. |\n| \"User did not specify `--mode` \u002F `--granularity`, `human` is the safe default\" | Always ask. Wrong mode cascades into every task's size and bundling. |\n| \"Phase boundary is fuzzy, I'll let SEQUENCE express the order\" | Phase boundaries are **hard gates**. If only some Phase N tasks must finish before a Phase N+1 task starts, that task belongs in Phase N. |\n| \"Ownership is just a hint, the implementer will figure it out\" | Ownership is an **operational edit boundary**. Vague Ownership = parallel-execution conflicts later. |\n| \"Spec Reference is empty for this task, I'll skip the section\" | Never omit. Use `N\u002FA — housekeeping\u002Frefactor only` when there is no spec. Empty section = ambiguity. |\n| \"Library introduction is part of the feature task, no need to split\" | Library introduction is its own task — in every mode. Safety invariant. |\n| \"20+ tasks in one set is fine if they are small\" | At >20, suggest splitting the spec. A task set that does not fit in human review will not be reviewed. |\n\n**Violating the letter of these rules is violating the spirit.** Safety invariants (DB migration separation, library introduction separation, phase hard gates) have no exceptions.\n\n## Arguments\n| Argument | Default | Description |\n|---|---|---|\n| `--lang \u003Clanguage>` | _(resolved or asked)_ | Output language for task documents: `ko` \\| `ja` \\| `en` \\| `es` \\| `zh`. Backward-compatible aliases: `korean` \\| `japanese` \\| `english` \\| `spanish` \\| `chinese`. |\n| `--mode \u003Cmode>` | _(asked)_ | Canonical task granularity option: `human` \\| `llm`. |\n| `--granularity \u003Cmode>` | _(alias)_ | Backward-compatible alias for `--mode`; accept `human` \\| `llm`. If both flags are present and conflict, stop and ask the user to resolve the conflict. |\n| `--tasks-dir \u003Cpath>` | `tasks\u002F` | Root directory where task directories are written. Override to support re-plan iteration in a separate directory (e.g., `--tasks-dir tasks-v2\u002F`). |\n| `--initials \u003Cvalue>` | _(resolved or asked)_ | Collaborator namespace for new task phases. Must match `^[a-z0-9]{2,4}$`; explicit invalid values are rejected. |\n| `--spec \u003Cpath>` | required for preview flags; optional for direct generation | Project-relative path for persisted previews and approved consumption. Direct generation retains Step 1 inputs, including inline specs. Preview specs must stay under `docs\u002F` unless the project defines another safe root explicitly. |\n| `--preview-only` | off | Persist the canonical preview and its reservation identity\u002Fref. Do not write task directories or `dependency-graph.md`. |\n| `--preview-path \u003Cpath>` | `docs\u002Fywc-plans\u002F\u003Cslug>.task-preview.md` | Persisted preview artifact path. Must be repository-relative Markdown under a safe root. |\n| `--approve-preview` | off | Consume a previously approved persisted preview with matching identity. Never re-decompose on this path. |\n| `--non-interactive` | off | Required when an autonomous caller consumes a previously approved preview without interactive confirmation. |\n| `--manifest \u003Cpath>` | off | Optional repository-relative architecture contract; omitted input permits root-only discovery. |\n| `--architecture-evidence \u003Cpath>` | off | Optional normalized evidence artifact; paired with `--manifest`, never raw evidence text. |\n\n## Language Option\nWhen `--lang` is not specified, resolve task document language using the shared Codex YWC policy:\n[`..\u002Freferences\u002Flanguage-resolution.md`](..\u002Freferences\u002Flanguage-resolution.md). The order is explicit `--lang` > project `.codex\u002Fywc.json` > project guidance (`AGENTS.md`, `CODEX.md`, `CLAUDE.md`) > user `~\u002F.codex\u002Fywc.json` > ask user.\n\nThis skill supports `ko` | `ja` | `en` | `es` | `zh` and backward-compatible aliases `korean` | `japanese` | `english` | `spanish` | `chinese`. Treat `zh` \u002F `chinese` as Simplified Chinese unless the user explicitly asks for another Chinese locale. There is no skill-level output-language default; for non-English output rules, read [references\u002Flanguage-policy.md](references\u002Flanguage-policy.md).\n\n### Collaborator Initials Gate\nRead [references\u002Fcollaborator-initials.md](references\u002Fcollaborator-initials.md)\nand resolve validated initials before any scan, compaction, reservation, or\nartifact write. Use explicit value, project config, user config, then one\ninteractive Git-identity confirmation; malformed tiers are skipped. Unresolved\nnon-interactive generation returns `NEEDS_CONTEXT` without downstream work.\n\n---\n\n## Preview Approval Gate\nFor persisted previews and approved-preview writes, `--spec` must resolve inside a project-declared safe spec root; absent a project declaration, use `docs\u002F`. `--preview-only` and `--approve-preview` are mutually exclusive. The preview identity binds the reserved phase and reservation ref; see [wide-refactor-decomposition.md](references\u002Fwide-refactor-decomposition.md).\n`--preview-only` returns `DONE` with its path\u002Frevision\u002Fdigest and phase reservation for a separate caller approval step; it writes no task directory or graph.\n`--approve-preview` consumes only an explicitly approved preview whose spec,\ntasks dir, language, mode, path, revision, digest, and phase reservation all\nmatch; under the lock, verify and reuse that reservation without selecting or\ncreating another phase. It never re-decomposes or refreshes the preview. Missing\napproval, identity mismatch, or unsafe paths return `NEEDS_CONTEXT` before task writes. A matching already-approved unchanged preview consumes once without another confirmation. Newly generated previews remain unapproved. Without\neither preview flag, preserve the existing direct-generation path without\nrequiring preview approval.\n\nFor the canonical preview identity fields, wide-refactor digest invalidation\nrules (`Refactor Phase`, `Batch ID`, `Depends On`), and the consume-only\nguarantee, **read\n[references\u002Fwide-refactor-decomposition.md](references\u002Fwide-refactor-decomposition.md)**.\n\n---\n\n## Task Design Principles\n\n### 1. Reviewability\n- Each task should be small enough for the intended executor to hold the full context in one session\n- Size guideline depends on the selected **Granularity Mode** (see Step 5):\n  - `human` mode: ~15 files \u002F ~500 LOC (optimized for a single-reviewer PR)\n  - `llm` mode: ~35 files \u002F ~1,200 LOC (optimized for an LLM agent completing a vertical slice in one run)\n- See [references\u002Fgranularity-modes.md](references\u002Fgranularity-modes.md) for the full rule set\n- Treat these numeric limits as advisory reviewability guidance, never as automatic authorization to bundle work; split tasks that significantly exceed the selected mode's guideline or cross a safety, feature, ownership, or phase boundary\n- In `llm` mode, bundling is limited to one feature with exclusive Ownership and explicitly declared Shared Surfaces\n- Prefer self-contained change units that leave the codebase buildable on completion\n\n### 2. Dependency Safety\n- Tasks must be strictly ordered by dependency\n- Earlier tasks must not depend on later tasks\n- Every task must be independently implementable at its position\n- If Task B depends on Task A, Task A must fully enable Task B\n- After each task completes, the codebase must remain consistent and buildable\n\n### 3. Database Migration Separation (Safety Invariant)\n- Database migrations must always be separated into their own task — in every mode\n- Include schema\u002Fmodel definitions\n- Never mix migration and feature implementation\n- Point the migration task's `task.md` checklist at the shared schema guide so the implementer applies the eight invariants that fail deterministically when omitted: [..\u002Freferences\u002Fschema\u002Fcore.md](..\u002Freferences\u002Fschema\u002Fcore.md) plus the stack file matching the project (`prisma.md` \u002F `sql-ddl.md` \u002F `drizzle.md` \u002F `typeorm.md`)\n\n### 4. Library Introduction Separation (Safety Invariant)\n- Create a dedicated task when introducing a new library\u002Fframework — in every mode\n- Never mix library introduction and feature implementation\n\n---\n\n## Workflow\n### Step 1: Verify Input Specification\nReceive the specification from the user. It may come in one of these forms:\n\n- Markdown document (file path or inline)\n- PRD (Product Requirements Document)\n- Feature requirements description\n- Change requirements for existing code\n\nWhen `--preview-only` or `--approve-preview` is in play, the input spec must\ncome from `--spec \u003Cpath>`. Persisted preview and approved consume modes do not\naccept an implicit inline spec because the identity must be replay-safe and\npath-bound.\n\nIf the specification is unclear, ask specific questions to clarify the scope.\n\n### Step 2: Resolve Initials and Collect Project Context\n\nGather project context for task verification commands, ownership paths, and category selection.\n\n**Targets to collect:**\n- `AGENTS.md`, `CODEX.md`, `CLAUDE.md` — project rules, language policy, CI commands. Prefer Codex-native instructions (`AGENTS.md` \u002F `CODEX.md`) when they conflict with Claude-only guidance.\n- `package.json`, `Makefile`, `pyproject.toml`, etc. — actual lint\u002Ftest\u002Fbuild commands\n- Project directory structure — `src\u002F` layout, monorepo status, existing module placement\n- Resolve `--tasks-dir` as a repository-relative path; reject absolute, escaping, or symlink-escaping paths with `NEEDS_CONTEXT`.\n- Resolve the default `tasks\u002F` from the repository root. The default is independent of `--spec` location and the current subdirectory; never infer `\u003Ctasks-dir>` from the parent directory of `--spec`. Resolve the output directory once and use that same path for preview analysis, approved consumption, writes, and validation.\n- Existing resolved `\u003Ctasks-dir>` and `\u003Ctasks-dir>\u002Fdependency-graph.md` — whether tasks already exist, numbering collision check\n- Docker environment — whether commands require a `docker exec` prefix\n- `docs\u002Fubiquitous-language.md` (if it exists) — canonical domain terms and \"Synonyms to Avoid\"; task names, Implementation Steps, and Ownership paths must use canonical terms and never use synonym identifiers\n\n**When existing tasks are present:**\n- Determine the next initials-scoped PHASE by parsing `tasks\u002Fdependency-graph.md` first when it exists, then scanning **both** `tasks\u002F` and `tasks\u002Fcompleted\u002F`, plus corresponding linked-worktree sources. Completed tasks are moved out of `tasks\u002F` into `tasks\u002Fcompleted\u002F` by the executors (`ywc-sequential-executor` \u002F `ywc-parallel-executor`), so scanning `tasks\u002F` alone misses them and risks reusing a number that already exists. For resolved `yk`, only `yk-\u003Csix-digit-phase>-\u003Csequence>-...` candidates contribute to the maximum; legacy and other-initials IDs remain readable but do not advance `yk`. An empty owned scope starts at `yk-000001-010`. If the graph and directory scan disagree, continue with the union-based next owned PHASE and report the mismatch as a concern.\n- **Proactive compaction gate**: after resolving the starting number above, check the resolved `\u003Ctasks-dir>\u002Fdependency-graph.md` line count (`wc -l`; `\u003Ctasks-dir>` defaults to `tasks`). If it exceeds **300 lines**, run `python3 codex\u002Fskills\u002Fywc-task-generator\u002Fscripts\u002Fcompact-dependency-graph.py \u003Ctasks-dir>` before generating any new task. Skip compaction during `--preview-only`; defer it until the approved reservation is verified during consumption. Direct generation and approved consumption may compact before task writes. Report the before\u002Fafter line count; if it remains above 300, active\u002Fplanned work accounts for the size and no further action is needed. The gate only removes phases whose every task is in `\u003Ctasks-dir>\u002Fcompleted\u002F` and never touches a phase with outstanding or unresolvable work.\n- Enumerate `git worktree list --porcelain`; inspect the same relative path in graph, active, and completed sources, reporting inaccessible\u002Fmismatched paths.\n- Scope the union maximum to prefixed IDs matching validated initials; keep legacy IDs readable without claiming another namespace. Empty owned scope starts at `\u003Cinitials>-000001-010`, with sequence `010`.\n- For direct generation and `--preview-only`, hold one common-Git lock across scan, selection, reservation, and task\u002Fpreview writes. For `--approve-preview`, verify and reuse its digest-bound reservation under the lock; do not select or reserve another phase.\n- Identify dependency relationships with existing tasks, preferring `dependency-graph.md` when present, and reflect them in the new tasks' `Depends On`\n\n### Step 3: Spec Review\nReview the specification for completeness and verify that sufficient information exists for decomposition. The goal is not to demand a perfect spec, but to confirm that the key information needed to determine task boundaries is present.\n\n**When unclear areas are found:**\n- Ask the user specific questions before starting task decomposition\n- Frame questions as choices (\"Is it A or B?\") rather than vague observations (\"This part is unclear\")\n- If the spec is sufficiently clear, skip this step and proceed directly\n\n### Step 4: Confirm Language\n\nIf `--lang` is provided, skip this step. Otherwise, resolve language using\n[`..\u002Freferences\u002Flanguage-resolution.md`](..\u002Freferences\u002Flanguage-resolution.md):\nproject `.codex\u002Fywc.json`, then project guidance (`AGENTS.md`, `CODEX.md`,\n`CLAUDE.md`), then user `~\u002F.codex\u002Fywc.json`, then ask if unresolved.\n\n> \"Which language should the task documents be written in? (`ko`, `ja`, `en`, `es`, or `zh`)\"\n\n### Step 5: Confirm Granularity Mode\n\nIf `--mode` is provided, use it and skip this confirmation. If `--granularity` is provided instead, treat it as a backward-compatible alias for `--mode` and skip this confirmation. If both are provided with different values, stop and ask the user to choose one mode before decomposing tasks.\n\nIf neither flag is provided, **always ask** the user which granularity mode to apply. Do not silently default — the correct mode depends on who will execute the tasks.\n\n> \"Which mode should the tasks be generated in?\n> - `human` — reviewable units (~15 files \u002F ~500 LOC)\n> - `llm` — larger vertical slices optimized for a single LLM agent run (~35 files \u002F ~1,200 LOC)\"\n\n**Mode selection criteria** (share with the user if they are unsure):\n- Choose `human` when a person will implement and code-review each task in sequence\n- Choose `llm` when tasks will be executed autonomously by an LLM agent (Codex, etc.) in isolated worktrees, and reviewed in aggregate after completion\n\nSafety invariants — DB migration separation, Library introduction separation, Phase hard gate, post-task buildability, and single phase per task — apply in **both modes**. Only advisory size and safe internal bundling differ; mode selection never authorizes cross-feature or invariant bundling.\n\nSee [references\u002Fgranularity-modes.md](references\u002Fgranularity-modes.md) for the full mode specification.\n\n**Persist the decision.** After parsing `--mode` \u002F `--granularity` or receiving confirmation, record the selected mode in a re-checkable location (e.g., a top-line scratchpad note, or `write_memory(\"granularity_mode\", \u003Cvalue>)` when Serena MCP is available) and apply it consistently through Steps 6–9 (size, bundling, Ownership scope, Implementation Steps depth, test.md inclusion). Mode drift mid-generation produces inconsistent task directories.\n\n### Step 5.5: Resolve Architecture Contract Packet\n\nIf `--manifest` or `--architecture-evidence` is supplied, resolve the pair\nthrough the shared `..\u002Fscripts\u002Farchitecture-invariants.py` helper before task\ndecomposition. A supplied manifest is repository-relative; invalid or missing\ninput returns `NEEDS_CONTEXT` without discovery fallback. Evidence without a\nvalid manifest is `NEEDS_CONTEXT`. A valid manifest without evidence preserves\nthe existing task-generation flow and adds no packet. Derive the bounded\n`--changed-path` set only from paths already named by the specification and\ncandidate task Ownership; never scan the repository broadly.\n\nRecord only this sanitized packet in preview\u002Ftask metadata and checklist\ncontext:\n\n```text\nArchitecture Contract Packet:\n- contract_state: \u003CVALIDATED | N\u002FA — no architecture contract | NEEDS_CONTEXT>\n- component_ids: \u003Caffected component ids>\n- rule_ids: \u003Caffected rule ids>\n- verifier_requirement: v1 validation-only; no verifier execution\n- invariant_verdict: \u003CMAINTAINED | VIOLATED | N\u002FA | NEEDS_CONTEXT>\n- evidence_artifact_path: \u003Crepository-relative path or N\u002FA>\n```\nMapping is separate: `status` controls dispatch; `aggregate_verdict` becomes `invariant_verdict`; rule IDs, sanitized evidence paths, and artifact path remain distinct. Successful audits use `contract_state: VALIDATED`; no-manifest is `N\u002FA — no architecture contract`; `NEEDS_CONTEXT` is terminal before writing.\nThe successful helper integration must provide `component_ids` from its bounded changed-path mapping, `contract_state: VALIDATED`, and the normalized `.ywc-architecture-invariants-evidence.json` as `evidence_artifact_path`; missing fields are `NEEDS_CONTEXT`, never values inferred from raw inputs. `verifier_requirement` is a fixed constant rather than a helper output: every rendered packet must carry the literal `v1 validation-only; no verifier execution` value, and a packet missing that line is incomplete. The `## Architecture Contract Packet` sections in `references\u002FREADME.md.template` and `references\u002Ftask.md.template` are conditional blocks — render them only when a validated packet exists for that task; when no `--manifest` is supplied, or a valid manifest is supplied without evidence, omit the heading and every field instead of writing an empty or placeholder-filled section.\nDo not copy manifest\u002Fevidence contents, raw evidence, command-like fields,\ntranscripts, full diffs, or inferred edges into task artifacts. Propagate\n`NEEDS_CONTEXT` before writing any preview or task directory. Surface\n`VIOLATED` as a task-generation finding; `N\u002FA` and `MAINTAINED` permit the\nexisting decomposition. The packet does not elevate task-generator authority\nor authorize execution. For the conditional task-owned quality packet, read [references\u002Ftask-metadata-rules.md](references\u002Ftask-metadata-rules.md) and apply its exact Ownership, required IDs\u002Fdigests, bounded evidence paths, redaction, omission, and `NEEDS_CONTEXT` rules; Task Verify remains separate.\n\n### Step 6: Task Decomposition\n\nAnalyze the specification and decompose it into tasks following the Task Design Principles and the selected Granularity Mode.\n\n**Scale guidelines** (task counts scale down in `llm` mode due to vertical bundling):\n- **Small** (1–3 tasks expected): A single phase is sufficient. Do not force multiple phases.\n- **Medium** (4–15 tasks in `human` mode, ~3–8 in `llm` mode): Organize into 2–3 phases. This is the typical case.\n- **Large** (15+ tasks in `human` mode, 8+ in `llm` mode): Suggest to the user that the spec itself be split into multiple features. A single task set exceeding 20 tasks becomes difficult to manage.\n\n#### Planning Advisor (optional, Pattern C)\n\nFor **Medium** and **Large** specs, phase boundary decisions and task size splits benefit from frontier judgment. A wrong Phase boundary cascades into every subsequent dependency declaration and is expensive to undo once the task directories are generated.\n\nThis skill applies **Pattern C** from [advisor-pattern.md](..\u002Freferences\u002Fadvisor-pattern.md): a **single** upfront higher-capability advisor call before writing any task directories. The executor handles everything else — Task Naming, Directory Generation, Dependency Graph, and Final Validation.\n\n**When to invoke**:\n- Scale is Medium (4–15 tasks) or Large (15+ tasks), AND\n- At least one of: phase boundary is non-obvious, multiple DB migrations or library introductions compete for Phase 1, or the spec touches more than two concurrent feature areas.\n- **Skip for Small specs** (1–3 tasks) — a single-phase decomposition is obvious and does not benefit from frontier reasoning. Adding an advisor call here wastes budget.\n\n**How to invoke**: Use Codex subagent delegation when the current session exposes a delegation tool. Request one higher-capability advisor pass with the bounded payload below. If no delegation tool is available, run the same advisor checklist inline as a separate bounded pass and record the fallback in the final handoff. Do not use Claude Code-only `Task` fields, `subagent_type`, or explicit Claude model pins in the Codex bundle.\n\n- **Spec summary** — ≤20 lines distilled from your Spec Review in Step 3 (not the full spec).\n- **First-pass task list** — task name + one-line description for each candidate task from your Step 6 decomposition.\n- **Selected mode and matching guideline** — the selected `human` or `llm` mode and exactly one corresponding advisory: `human` → ~15 files \u002F ~500 LOC; `llm` → ~35 files \u002F ~1,200 LOC. Do not evaluate a task against the other mode's guideline.\n- **Known conflicts \u002F shared surfaces** — any Ownership overlaps or Shared Surfaces you identified.\n- **Project context essentials** — monorepo structure, existing phases in `tasks\u002F`, tech-stack constraints affecting dependency order.\n\nAsk the advisor for three things:\n\n1. **Phase boundary recommendations** — which tasks belong to which phase, and why each boundary is a hard gate.\n2. **Task size verification** — whether any task exceeds the selected mode's matching advisory guideline and, independently, whether a feature, Ownership, Shared Surface, safety-invariant, phase, or buildability boundary requires a split even when the numeric limit is met.\n3. **Dependency cycle risk** — any tasks with circular implicit dependencies the first-pass missed.\n\n**Budget**: exactly **1** advisor pass per invocation of this skill. Pattern C explicitly rules out re-invocation during execution — if the initial plan proves wrong, re-run the whole skill with refined input rather than calling another advisor mid-generation. This rule exists because mid-generation re-planning leads to inconsistent task directories; a fresh start is cleaner and more auditable.\n\n**Payload rules**: the summary and task list together must not exceed ~200 lines of payload. If the first-pass decomposition is already larger, narrow it before the advisor call (split the spec into sub-specs and plan each separately). Never forward the full spec verbatim.\n\n**Advisor output format** (≤300 words):\n- Phase boundary plan with one-line rationale per boundary\n- Task size concerns (if any)\n- Cycle\u002Fordering warnings (if any)\n- Single \"proceed\" or \"reconsider with refinements\" verdict\n\nAfter receiving the verdict, the executor either continues to Step 7 (Task Naming) with the adjusted plan, or surfaces the \"reconsider\" verdict to the user for refinement before proceeding.\n\n### Step 7: Task Naming\nEach new task name follows this format:\n\n```\n[INITIALS]-[PHASE]-[SEQUENCE]-[CATEGORY]-[SHORT-DESCRIPTION]\n```\n> Example: `yk-000001-010-db-create-user-table` = INITIALS `yk` + PHASE `000001` + SEQUENCE `010` + category + description.\n\n**Numbering Rules:**\n- PHASE: **6-digit** number (`000001`, `000002`, ...) — the wider width reserves headroom for multi-year project growth\n- INITIALS: validated **2–4 lowercase alphanumeric** characters resolved before allocation\n- SEQUENCE: 3-digit number (`010`, `020`, `030`, ...)\n- Sequence increments by 10 (allows inserting tasks later without renumbering)\n- Always use hyphen (`-`) to separate PHASE and SEQUENCE for readability\n\n**Category:**\n- `lib` — New library\u002Fframework introduction\n- `db` — Database migration, schema change\n- `api` — API endpoint implementation\n- `domain` — Business logic, service layer\n- `worker` — Background job, async processing\n- `ui` — Frontend component, page\n- `test` — Test-only task (e2e, load test, etc.)\n- `refactor` — Code structure improvement with no behavior change\n- `infra` — CI\u002FCD, deployment, configuration\n- `config` — Environment variables, feature flags, `.env` setup (distinct from infra: infra covers CI\u002FCD pipelines and deployment, config covers application-level settings)\n\n**Naming Quality Rules:**\n- Use only lowercase and hyphens\n- Keep names concise, descriptive, and filesystem-friendly\n- Express exactly one primary concern (in `llm` mode, \"one primary concern\" may cover a vertical slice of a single feature)\n- Avoid vague names\n- Do not use \"and\" in task names — use a unifying noun instead (e.g., `user-auth` rather than `registration-and-login`)\n\n**Examples:**\n- `yk-000001-010-db-create-user-table`\n- `yk-000001-020-lib-setup-auth-library`\n- `yk-000001-030-api-user-registration`\n- `yk-000002-010-ui-login-form`\n\nExisting numeric and unprefixed task IDs remain valid in dependencies, graph\nparsing, compaction, executor ranges, and completion moves. Do not rename\nlegacy directories or count them as another collaborator's prefixed namespace.\n\n### Step 8: Phase Organization\n- PHASE represents a meaningful dependency stage (e.g., `000001` = foundation, `000002` = core features, `000003` = integration)\n- Do not create phases arbitrarily — a new phase means \"all tasks in the previous phase must be complete before this phase can start\"\n- Tasks within the same phase may have internal ordering via SEQUENCE, but conceptually belong to the same stage\n\n**Phase Boundary Rules:**\n- Phase boundaries are **hard gates**. No task in Phase N+1 can start before all tasks in Phase N are complete\n- Dependencies between tasks within the same phase are allowed (order expressed via SEQUENCE)\n- If only some tasks in Phase N need to be complete before a specific Phase N+1 task can start, that task should be placed in Phase N instead\n- Criterion for splitting phases: \"Do all outputs from this stage need to be in place for the next stage's work to be meaningful?\"\n\n### Step 9: Directory and File Generation\nUse the pre-resolved `--tasks-dir` path for every task output. If omitted, use the repository-root `tasks\u002F`; never derive it from the spec's parent. With `--preview-only`, write the preview and its digest-bound reservation ref, but no task files or graph. With `--approve-preview`, create task files only after explicit approval and identity, reservation, authorization, and containment checks pass. Without preview flags, preserve direct generation after existing validation; no preview approval is required.\n\nGenerate the following structure for each task:\n\n```\ntasks\u002F[TASK_NAME]\u002F\n├── README.md\n├── task.md\n└── test.md (optional — included when manual verification is needed)\n```\n\nRefer to templates in the `references\u002F` directory when writing files:\n- `references\u002FREADME.md.template` — Task overview and dependency documentation\n- `references\u002Ftask.md.template` — Implementation checklist\n- `references\u002Ftest.md.template` — Manual test plan\n- `references\u002Fdependency-graph.md.template` — Top-level dependency summary\n\n#### README.md Core Elements\n\nEach README.md must include the following:\n\n- **Purpose**: What this task achieves\n- **Scope**: What is included\n- **Spec Reference**: Specification\u002Fdesign documents that must be consulted before implementation (see \"Spec Reference Rules\" below)\n- **Dependencies**: Bidirectional dependency tracking\n  - **Depends On**: Predecessor tasks this task depends on (specify what each provides)\n  - **Depended By**: Successor tasks that depend on this task (specify what each needs)\n- **Key Files**: List of files expected to be created or modified\n- **Notes**: Important considerations, design decisions\n- **Out of Scope**: What is intentionally excluded\n- **Parallel Execution Metadata**:\n  - **Model Hint**: Exactly one advisory tier and evidence-based rationale per task; use [task-metadata-rules.md](references\u002Ftask-metadata-rules.md) as the canonical tier definition.\n  - **Ownership**: Files, directories, modules, contracts this task may modify\n  - **Shared Surfaces**: Shared boundaries that may cause merge or design conflicts\n  - **Conflicts With**: Tasks that must not run in parallel\n  - **Parallelizable After**: Minimum merged baseline before isolated execution can safely start\n  - **Task Verify**: Exact commands that prove this task is complete\n\nBidirectional dependency tracking matters because developers executing a task can immediately see which downstream tasks are affected, allowing them to consider downstream requirements when designing interfaces.\n\n#### Spec Reference Rules\n\nEvery task's README.md must include a `## Spec Reference` section with three subsections: **Primary Sources** (links to source-of-truth docs, project-relative by default), **Summary** (2–5 sentence orientation), and **Out of Scope (from spec)** (scope-creep guardrail).\n\n**Hard rule**: never leave Spec Reference empty or omitted — use `N\u002FA — no external spec (housekeeping \u002F refactor \u002F config only)` when no spec exists.\n\nFor the full structure, the external-URL policy, the link-vs-summary drift protocol, and the cross-task scope-handoff convention, **read [references\u002Fspec-reference-rules.md](references\u002Fspec-reference-rules.md)**.\n\n#### Parallel Execution Metadata Rules\n\nRead [references\u002Ftask-metadata-rules.md](references\u002Ftask-metadata-rules.md) for advisory tier selection and its runtime boundaries. Every task requires a Model Hint, including tasks that will not run in parallel.\n\nWhen tasks may be executed via `git worktree`, Codex, or other isolated worker setups, every task's README.md must declare:\n\n- **Ownership** — concrete edit boundary (file paths, globs, modules, contracts). Bad: `backend`, `frontend related files`. Good: `api\u002Fsrc\u002Froutes\u002Fmetrics.ts`, `OpenAPI contract: \u002Fapi\u002Fv1\u002Fmetrics`.\n- **Shared Surfaces** — shared boundaries (DB schema, API contracts, event payloads, workspace config, CI config) that may cause conflicts even without direct file overlap.\n- **Conflicts With** — tasks that must not run in parallel even when dependencies allow. Write `(None identified)` if none.\n- **Parallelizable After** — the minimum merged baseline for safely starting isolated execution (often, but not always, the same as `Depends On`).\n- **Task Verify** — task-specific commands that prove completion (avoid project-wide gates only).\n\nFor the full rule set including good\u002Fbad Ownership examples, the complete shared-surfaces taxonomy, and the dependency-graph Parallel Execution Notes format, **read [references\u002Fparallel-execution-metadata.md](references\u002Fparallel-execution-metadata.md)**.\n\n#### task.md Core Elements\n\nEach task.md must include the following:\n\n- **Prerequisites**: Predecessor task completion status to verify before starting (checkbox format)\n- **Allowed Edit Scope**: A brief restatement of the `README.md` Ownership as an operational boundary\n- **Stop Conditions**: Conditions under which the implementer should stop and report rather than continuing\n- **Implementation Steps**: Specific implementation steps (checkbox format)\n  - Each step must reference specific files, modules, functions, or behaviors\n  - Do not use generic placeholders like \"implement core logic\" or \"handle edge cases\"\n  - Example: `Create src\u002Fmodels\u002Fuser.ts with User entity definition`\n- **Task Verify**: Task-specific verification command checklist\n- **Verification**: Confirm lint, typecheck, test, and build pass (use the project's actual commands)\nFor a `ui` task changing end-user layout\u002Fstyles, link to [`mobile-first-ui.md`](..\u002Freferences\u002Fmobile-first-ui.md) and order Implementation Steps narrowest-viewport base first, then outward `min-width` expansion. Require a PC\u002Ftablet-only exception only when explicitly stated; do not apply this to logic-only UI, admin\u002Finternal, non-UI, or unchanged legacy desktop-first CSS (no retrofit).\n**Duplicate-sensitive write Task Verify rule**: When a task touches payment, order creation, provisioning, credit \u002F balance \u002F stock \u002F quota mutation, or another duplicate-sensitive write flow, its `task.md` must include concrete Task Verify commands or test scenarios for concurrent request behavior, transaction rollback or equivalent consistency rollback on partial failure, and idempotency retry \u002F double-click behavior. The task must name the selected mechanism where applicable: atomic conditional update, row lock, optimistic lock, idempotency key, unique constraint, or persisted request\u002Fresult record. Expected results must be observable: lock\u002Fversion conflict response, exhausted stock\u002Fbalance\u002Fquota response, rollback state after a forced mid-flow failure, and duplicate retry returning the prior result without repeating the side effect. If the project has no practical local concurrency harness, the task may record a named exception only with replacement verification such as a code-level lock\u002Ftransaction proof or an integration-test plan.\n\n#### test.md (optional)\n\nWrite structured scenario-based tests (Steps + Expected Result).\n\n**Include when:**\n- UI component or screen behavior verification → always include\n- External service integration (S3, Email, Payment, etc.) → always include\n- Browser-dependent features → always include\n- Complex user workflows (multi-step) → recommended\n\n**May omit when:**\n- Pure API endpoints (coverable by unit + integration tests)\n- Database migrations (schema application verified via task.md Verification)\n- Library introduction tasks (import and basic operation verified via Verification)\n\n### Step 10: Generate Dependency Graph\nAfter generating tasks in direct-generation or approved-consumption mode, create `\u003Ctasks-dir>\u002Fdependency-graph.md` at the top level (where `\u003Ctasks-dir>` is the value of `--tasks-dir`, defaulting to `tasks\u002F`). This file serves as the single source of truth for execution order. Preview-only stops after writing and validating its preview artifact; it does not create or modify the dependency graph.\n\nRefer to `references\u002Fdependency-graph.md.template` for format. List tasks by phase and express each task's dependencies using arrow notation.\nThis graph must be consistent with the Dependencies sections in individual README.md files.\nOnce every task in a phase reaches `\u003Ctasks-dir>\u002Fcompleted\u002F`, the shared completion marker automatically compacts that phase (see `codex\u002Fskills\u002Fywc-task-generator\u002Fscripts\u002Fcompact-dependency-graph.py`) — no special markup is required; the script reads the `## Phase NNNNNN`, `## Parallel Execution Notes`, and `## Visual Dependency Graph` headings directly.\n### Step 11: Final Validation\nBefore returning a preview, direct task set, or consumed task set, apply [requirement coverage and output validation](references\u002Foutput-and-validation.md), including its attributable coverage summary and interface\u002Fverification mapping. A clean preview returns `DONE` with its preview identity; approval is a separate caller step. Then verify the following, applying task-artifact checks only to direct generation or approved consumption:\n\n**Dependency & Structure:**\n- [ ] No forward dependencies (earlier tasks do not depend on later tasks)\n- [ ] No circular dependencies\n- [ ] Phase boundaries correctly set as hard gates\n- [ ] Database migrations and library introductions separated into their own tasks (Safety Invariant in both modes)\n- [ ] No numbering collisions across the resolved graph, active\u002Fcompleted tasks, and every linked worktree source\n- [ ] Graph\u002Fdirectory numbering mismatches reported as concerns (if any)\n**Naming & Size:**\n- [ ] Naming convention followed (`[INITIALS:2-4]-[PHASE:6]-[SEQUENCE:3]-[CATEGORY]-[SHORT-DESCRIPTION]`)\n- [ ] Direct\u002Fpreview generation reserves under the common-Git lock; approved consumption verifies and reuses the preview's exact phase reservation under that lock without selecting another phase.\n- [ ] PHASE is 6 digits, SEQUENCE is 3 digits, separated by hyphen\n- [ ] No task name contains \"and\" (single concern check)\n- [ ] Each task fits within the selected Granularity Mode's size guideline\n- [ ] Selected Granularity Mode applied consistently (bundling, Implementation Steps depth, Ownership scope)\n**README.md Quality:**\n- [ ] Both \"Depends On\" and \"Depended By\" included\n- [ ] Parallel Execution Metadata included (Model Hint, Ownership, Shared Surfaces, Conflicts With, Parallelizable After, Task Verify)\n- [ ] Ownership is narrow enough to serve as an actual edit boundary, not just a restatement of Key Files\n- [ ] **Spec Reference** section included — Primary Sources, Summary, and Out of Scope (from spec) all filled in or explicitly marked `N\u002FA`\n- [ ] If Primary Sources include external URLs, only if the project allows it (default is project-relative paths only)\n\n**task.md Quality:**\n- [ ] Specific Implementation Steps included (no generic placeholders)\n- [ ] Prerequisites section included\n- [ ] Allowed Edit Scope and Stop Conditions included\n- [ ] Verification commands match the project's actual commands (based on context collected in Step 2)\n- [ ] Duplicate-sensitive write tasks include concrete concurrent request, rollback, and idempotency retry verification or a named no-harness exception with replacement evidence\n\n**Consistency:**\n- [ ] Every task has a valid advisory Model Hint and rationale; preview, README, and final summary agree before writes.\n- [ ] Dependency Graph Summary matches individual README.md files\n- [ ] If the task set is intended for parallel execution, Dependency Graph includes Parallel Execution Notes\n- [ ] Codebase is buildable after each task completes\n- [ ] Selected language's writing rules are followed\n\n---\n\n## Output Format\n\nThe final output includes:\n\n1. **Task list summary** organized by phase with Task ID, Title, and Model Hint (tier plus rationale) for every task, matching the preview when used and every generated README\n2. **Directory generation** under `\u003Ctasks-dir>\u002F`\n3. **Dependency Graph** at `\u003Ctasks-dir>\u002Fdependency-graph.md`, including Parallel Execution Notes when needed\n\nPreview-only returns the task summary and persisted preview identity; it does not claim that task directories or a dependency graph were generated. Items 2 and 3 apply only to direct generation or approved consumption.\n\nWhen parallel execution is expected, verify that each task is safe for isolated worktree or agent execution.\n\n---\n\n## Validation\n\nFor allocation, run:\n```bash\nRESOLVER_LAUNCHER=\"$(bash \"${CODEX_HOME:-$HOME\u002F.codex}\u002Fskills\u002Fscripts\u002Fselect-resolver-launcher.sh\" 2>&1)\" || { echo \"BLOCKED: ${RESOLVER_LAUNCHER#BLOCKED: }\" >&2; exit 3; }\nINITIALS_TEST=\"$(bash \"$RESOLVER_LAUNCHER\" \"ywc-task-generator\u002Fscripts\u002Ftest-initials-allocation.sh\" bash)\" || exit $?\nPARSER_TEST=\"$(bash \"$RESOLVER_LAUNCHER\" \"ywc-task-generator\u002Fscripts\u002Ftest-parser-ids.sh\" bash)\" || exit $?\nbash \"$INITIALS_TEST\" && bash \"$PARSER_TEST\"\n```\nTask Verify must cover config precedence\u002Fmalformed tiers, linked-worktree\ngraph\u002Factive\u002Fcompleted sources, scoped empty-graph maxima, concurrent distinct\nreservations, and missing-initials `NEEDS_CONTEXT` before any write.\n\nBefore returning `DONE`, verify:\n\n- [ ] The spec or plan source was read and each task traces to a source requirement.\n- [ ] For preview-only, verify the persisted preview artifact and its identity, and verify that no task directory or dependency graph was created or modified.\n- [ ] For direct generation or approved consumption, verify every emitted task directory contains `README.md` and `task.md`, and `\u003Ctasks-dir>\u002Fdependency-graph.md` exists. All task artifacts must be under the pre-resolved `\u003Ctasks-dir>`; never beside the spec unless explicitly selected.\n- [ ] UI or behavior-changing tasks have a `test.md` or an explicit test exception.\n- [ ] For direct generation or approved consumption, `dependency-graph.md` exists and matches the generated task directories.\n- [ ] Parallel Execution Notes are present when tasks are intended for parallel execution.\n- [ ] The selected output language matches `--lang` or the shared YWC language resolution result.\n- [ ] The selected Granularity Mode matches `--mode`, `--granularity`, or the user's explicit confirmation.\n- [ ] The final response reports the generated directory, task count, graph path, and `Status: DONE | DONE_WITH_CONCERNS | BLOCKED | NEEDS_CONTEXT`.\n\n## Task Execution Convention\n\nFor execution rules after task generation (completion handling, directory moves, etc.), see `references\u002Fexecution-convention.md`.\n",{"name":5,"description":7},"main","7e8ac872e53df8c89d75663e031626281a2a2f5c","codex\u002Fskills\u002Fywc-task-generator",1,"https:\u002F\u002Fgithub.com\u002Fyongwoon\u002Fywc-agent-toolkit","https:\u002F\u002Fgithub.com\u002Fyongwoon","https:\u002F\u002Fwww.webhani.com","2026-11-08 17:14:59",{"source":8,"skillName":5,"commands":32,"prompt":42,"zipAvailable":43},[33,36,39],{"tool":34,"command":35},"npx","npx skills add yongwoon\u002Fywc-agent-toolkit --skill ywc-task-generator",{"tool":37,"command":38},"bunx","bunx skills add yongwoon\u002Fywc-agent-toolkit --skill ywc-task-generator",{"tool":40,"command":41},"pnpm","pnpm dlx skills add yongwoon\u002Fywc-agent-toolkit --skill ywc-task-generator","安装技能，技能地址：https:\u002F\u002Fgithub.com\u002Fyongwoon\u002Fywc-agent-toolkit\u002Fblob\u002Fmain\u002Fcodex\u002Fskills\u002Fywc-task-generator\u002FSKILL.md",true,[45,51,55,59,63,67,71,75,79,83,87,91,95,99,103,107,111,115,119,123,127,131,137,141],{"path":46,"resourceType":47,"resourceTypeDesc":48,"sizeBytes":49,"sourceHtmlUrl":50},"agents\u002Fopenai.yaml","REFERENCE","参考文档",497,"https:\u002F\u002Fgithub.com\u002Fyongwoon\u002Fywc-agent-toolkit\u002Fblob\u002Fmain\u002Fcodex\u002Fskills\u002Fywc-task-generator\u002Fagents\u002Fopenai.yaml",{"path":52,"resourceType":47,"resourceTypeDesc":48,"sizeBytes":53,"sourceHtmlUrl":54},"evals\u002Fevals.json",34543,"https:\u002F\u002Fgithub.com\u002Fyongwoon\u002Fywc-agent-toolkit\u002Fblob\u002Fmain\u002Fcodex\u002Fskills\u002Fywc-task-generator\u002Fevals\u002Fevals.json",{"path":56,"resourceType":47,"resourceTypeDesc":48,"sizeBytes":57,"sourceHtmlUrl":58},"README.en.md",7985,"https:\u002F\u002Fgithub.com\u002Fyongwoon\u002Fywc-agent-toolkit\u002Fblob\u002Fmain\u002Fcodex\u002Fskills\u002Fywc-task-generator\u002FREADME.en.md",{"path":60,"resourceType":47,"resourceTypeDesc":48,"sizeBytes":61,"sourceHtmlUrl":62},"README.es.md",9041,"https:\u002F\u002Fgithub.com\u002Fyongwoon\u002Fywc-agent-toolkit\u002Fblob\u002Fmain\u002Fcodex\u002Fskills\u002Fywc-task-generator\u002FREADME.es.md",{"path":64,"resourceType":47,"resourceTypeDesc":48,"sizeBytes":65,"sourceHtmlUrl":66},"README.ja.md",8576,"https:\u002F\u002Fgithub.com\u002Fyongwoon\u002Fywc-agent-toolkit\u002Fblob\u002Fmain\u002Fcodex\u002Fskills\u002Fywc-task-generator\u002FREADME.ja.md",{"path":68,"resourceType":47,"resourceTypeDesc":48,"sizeBytes":69,"sourceHtmlUrl":70},"README.ko.md",8819,"https:\u002F\u002Fgithub.com\u002Fyongwoon\u002Fywc-agent-toolkit\u002Fblob\u002Fmain\u002Fcodex\u002Fskills\u002Fywc-task-generator\u002FREADME.ko.md",{"path":72,"resourceType":47,"resourceTypeDesc":48,"sizeBytes":73,"sourceHtmlUrl":74},"README.md",10495,"https:\u002F\u002Fgithub.com\u002Fyongwoon\u002Fywc-agent-toolkit\u002Fblob\u002Fmain\u002Fcodex\u002Fskills\u002Fywc-task-generator\u002FREADME.md",{"path":76,"resourceType":47,"resourceTypeDesc":48,"sizeBytes":77,"sourceHtmlUrl":78},"README.zh.md",7894,"https:\u002F\u002Fgithub.com\u002Fyongwoon\u002Fywc-agent-toolkit\u002Fblob\u002Fmain\u002Fcodex\u002Fskills\u002Fywc-task-generator\u002FREADME.zh.md",{"path":80,"resourceType":47,"resourceTypeDesc":48,"sizeBytes":81,"sourceHtmlUrl":82},"references\u002Fcollaborator-initials.md",7430,"https:\u002F\u002Fgithub.com\u002Fyongwoon\u002Fywc-agent-toolkit\u002Fblob\u002Fmain\u002Fcodex\u002Fskills\u002Fywc-task-generator\u002Freferences\u002Fcollaborator-initials.md",{"path":84,"resourceType":47,"resourceTypeDesc":48,"sizeBytes":85,"sourceHtmlUrl":86},"references\u002Fdependency-graph.md.template",2848,"https:\u002F\u002Fgithub.com\u002Fyongwoon\u002Fywc-agent-toolkit\u002Fblob\u002Fmain\u002Fcodex\u002Fskills\u002Fywc-task-generator\u002Freferences\u002Fdependency-graph.md.template",{"path":88,"resourceType":47,"resourceTypeDesc":48,"sizeBytes":89,"sourceHtmlUrl":90},"references\u002Fexecution-convention.md",1756,"https:\u002F\u002Fgithub.com\u002Fyongwoon\u002Fywc-agent-toolkit\u002Fblob\u002Fmain\u002Fcodex\u002Fskills\u002Fywc-task-generator\u002Freferences\u002Fexecution-convention.md",{"path":92,"resourceType":47,"resourceTypeDesc":48,"sizeBytes":93,"sourceHtmlUrl":94},"references\u002Fgranularity-modes.md",5675,"https:\u002F\u002Fgithub.com\u002Fyongwoon\u002Fywc-agent-toolkit\u002Fblob\u002Fmain\u002Fcodex\u002Fskills\u002Fywc-task-generator\u002Freferences\u002Fgranularity-modes.md",{"path":96,"resourceType":47,"resourceTypeDesc":48,"sizeBytes":97,"sourceHtmlUrl":98},"references\u002Flanguage-policy.md",2975,"https:\u002F\u002Fgithub.com\u002Fyongwoon\u002Fywc-agent-toolkit\u002Fblob\u002Fmain\u002Fcodex\u002Fskills\u002Fywc-task-generator\u002Freferences\u002Flanguage-policy.md",{"path":100,"resourceType":47,"resourceTypeDesc":48,"sizeBytes":101,"sourceHtmlUrl":102},"references\u002Foutput-and-validation.md",3822,"https:\u002F\u002Fgithub.com\u002Fyongwoon\u002Fywc-agent-toolkit\u002Fblob\u002Fmain\u002Fcodex\u002Fskills\u002Fywc-task-generator\u002Freferences\u002Foutput-and-validation.md",{"path":104,"resourceType":47,"resourceTypeDesc":48,"sizeBytes":105,"sourceHtmlUrl":106},"references\u002Fparallel-execution-metadata.md",2559,"https:\u002F\u002Fgithub.com\u002Fyongwoon\u002Fywc-agent-toolkit\u002Fblob\u002Fmain\u002Fcodex\u002Fskills\u002Fywc-task-generator\u002Freferences\u002Fparallel-execution-metadata.md",{"path":108,"resourceType":47,"resourceTypeDesc":48,"sizeBytes":109,"sourceHtmlUrl":110},"references\u002FREADME.md.template",9380,"https:\u002F\u002Fgithub.com\u002Fyongwoon\u002Fywc-agent-toolkit\u002Fblob\u002Fmain\u002Fcodex\u002Fskills\u002Fywc-task-generator\u002Freferences\u002FREADME.md.template",{"path":112,"resourceType":47,"resourceTypeDesc":48,"sizeBytes":113,"sourceHtmlUrl":114},"references\u002Fspec-reference-rules.md",1844,"https:\u002F\u002Fgithub.com\u002Fyongwoon\u002Fywc-agent-toolkit\u002Fblob\u002Fmain\u002Fcodex\u002Fskills\u002Fywc-task-generator\u002Freferences\u002Fspec-reference-rules.md",{"path":116,"resourceType":47,"resourceTypeDesc":48,"sizeBytes":117,"sourceHtmlUrl":118},"references\u002Ftask-metadata-rules.md",3853,"https:\u002F\u002Fgithub.com\u002Fyongwoon\u002Fywc-agent-toolkit\u002Fblob\u002Fmain\u002Fcodex\u002Fskills\u002Fywc-task-generator\u002Freferences\u002Ftask-metadata-rules.md",{"path":120,"resourceType":47,"resourceTypeDesc":48,"sizeBytes":121,"sourceHtmlUrl":122},"references\u002Ftask.md.template",6687,"https:\u002F\u002Fgithub.com\u002Fyongwoon\u002Fywc-agent-toolkit\u002Fblob\u002Fmain\u002Fcodex\u002Fskills\u002Fywc-task-generator\u002Freferences\u002Ftask.md.template",{"path":124,"resourceType":47,"resourceTypeDesc":48,"sizeBytes":125,"sourceHtmlUrl":126},"references\u002Ftest.md.template",382,"https:\u002F\u002Fgithub.com\u002Fyongwoon\u002Fywc-agent-toolkit\u002Fblob\u002Fmain\u002Fcodex\u002Fskills\u002Fywc-task-generator\u002Freferences\u002Ftest.md.template",{"path":128,"resourceType":47,"resourceTypeDesc":48,"sizeBytes":129,"sourceHtmlUrl":130},"references\u002Fwide-refactor-decomposition.md",3371,"https:\u002F\u002Fgithub.com\u002Fyongwoon\u002Fywc-agent-toolkit\u002Fblob\u002Fmain\u002Fcodex\u002Fskills\u002Fywc-task-generator\u002Freferences\u002Fwide-refactor-decomposition.md",{"path":132,"resourceType":133,"resourceTypeDesc":134,"sizeBytes":135,"sourceHtmlUrl":136},"scripts\u002Fcompact-dependency-graph.py","SCRIPT","脚本",6697,"https:\u002F\u002Fgithub.com\u002Fyongwoon\u002Fywc-agent-toolkit\u002Fblob\u002Fmain\u002Fcodex\u002Fskills\u002Fywc-task-generator\u002Fscripts\u002Fcompact-dependency-graph.py",{"path":138,"resourceType":133,"resourceTypeDesc":134,"sizeBytes":139,"sourceHtmlUrl":140},"scripts\u002Ftest-initials-allocation.sh",8755,"https:\u002F\u002Fgithub.com\u002Fyongwoon\u002Fywc-agent-toolkit\u002Fblob\u002Fmain\u002Fcodex\u002Fskills\u002Fywc-task-generator\u002Fscripts\u002Ftest-initials-allocation.sh",{"path":142,"resourceType":133,"resourceTypeDesc":134,"sizeBytes":143,"sourceHtmlUrl":144},"scripts\u002Ftest-parser-ids.sh",2239,"https:\u002F\u002Fgithub.com\u002Fyongwoon\u002Fywc-agent-toolkit\u002Fblob\u002Fmain\u002Fcodex\u002Fskills\u002Fywc-task-generator\u002Fscripts\u002Ftest-parser-ids.sh",[146,152,158,165,172,179],{"skillNo":147,"name":148,"nameZh":6,"description":6,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":9,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":149,"categories":150,"keywords":151,"publishedAt":18},"01M72TQ84GTCPCV2VJ778V17TW","missing-frontmatter","https:\u002F\u002Fgithub.com\u002Fyongwoon\u002Fywc-agent-toolkit\u002Fblob\u002Fmain\u002Fscripts\u002Ffixtures\u002Fcodex-description-limit\u002Fmissing-frontmatter\u002FSKILL.md",[],[],{"skillNo":153,"name":154,"nameZh":6,"description":6,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":9,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":155,"categories":156,"keywords":157,"publishedAt":18},"01M72TQ6M4BFWE7P9EPW0GZDCM","malformed-frontmatter","https:\u002F\u002Fgithub.com\u002Fyongwoon\u002Fywc-agent-toolkit\u002Fblob\u002Fmain\u002Fscripts\u002Ffixtures\u002Fcodex-description-limit\u002Fmalformed-frontmatter\u002FSKILL.md",[],[],{"skillNo":159,"name":160,"nameZh":6,"description":161,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":9,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":162,"categories":163,"keywords":164,"publishedAt":18},"01M72TQ3ESGKXZ4QRWGTX92BKJ","fixture-501-fail","(ywc) Use when aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa Do not use for x.","https:\u002F\u002Fgithub.com\u002Fyongwoon\u002Fywc-agent-toolkit\u002Fblob\u002Fmain\u002Fscripts\u002Ffixtures\u002Fcodex-description-limit\u002F501-fail\u002FSKILL.md",[],[],{"skillNo":166,"name":167,"nameZh":6,"description":168,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":9,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":169,"categories":170,"keywords":171,"publishedAt":18},"01M72TQ0HJ8H3R37J9WPWGKJZR","fixture-500-pass","(ywc) Use when aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa Do not use for x.","https:\u002F\u002Fgithub.com\u002Fyongwoon\u002Fywc-agent-toolkit\u002Fblob\u002Fmain\u002Fscripts\u002Ffixtures\u002Fcodex-description-limit\u002F500-pass\u002FSKILL.md",[],[],{"skillNo":173,"name":174,"nameZh":6,"description":175,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":9,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":176,"categories":177,"keywords":178,"publishedAt":18},"01M72TPWNCC7NEYD0D9N12KSQS","ywc-worktrees","(ywc) Use when creating, auditing, pruning, or resolving git worktrees for isolated task workspaces, whether delegated from an executor or used standalone. Triggers: \"worktree 생성\", \"worktree audit\", \"worktree 정리\", \"worktree prune\", \"ywc-worktrees\", \"create worktree\". Do not use for branch creation\u002Fdeletion alone, PR creation (use ywc-create-pr), CI or merge delivery (use ywc-finish-branch), task orchestration (use ywc-sequential-executor or ywc-parallel-executor), or non-git isolation.","https:\u002F\u002Fgithub.com\u002Fyongwoon\u002Fywc-agent-toolkit\u002Fblob\u002Fmain\u002Fplugins\u002Fywc-agent-toolkit\u002Fskills\u002Fywc-worktrees\u002FSKILL.md",[],[],{"skillNo":180,"name":181,"nameZh":6,"description":182,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":9,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":183,"categories":184,"keywords":185,"publishedAt":18},"01M72TPRNR3E6QWB91GM98ABBV","ywc-wayfinder","(ywc) Use when a large or uncertain change needs multi-session discovery with a deterministic local map and exactly one active ticket before planning or implementation. Triggers: \"wayfinder\", \"길찾기\", \"discovery map\", \"multi-session discovery\", \"탐색 ticket\", \"経路整理\", \"探索マップ\", \"ywc-wayfinder\". Do not use for ordinary planning (use ywc-plan), idea clarification (use ywc-brainstorm), direct implementation, or tracker\u002Fdatabase writes outside the local Markdown map.","https:\u002F\u002Fgithub.com\u002Fyongwoon\u002Fywc-agent-toolkit\u002Fblob\u002Fmain\u002Fplugins\u002Fywc-agent-toolkit\u002Fskills\u002Fywc-wayfinder\u002FSKILL.md",[],[]]