wb-flow 1.0.0-r01

package/templates/commands/wbGit/wbGit_template.md

@@ -0,0 +1,160 @@
+# /wbGit: Execution Template
+
+
+<!-- HELP_GATE_START -->
+## Help intercept (handle FIRST — before any other action)
+
+**If `$ARGUMENTS` contains `--help`, `-h`, or `--h`** (case-insensitive, anywhere in the args), DO NOT execute the command's normal procedure. Instead:
+
+1. Output the **HELP BLOCK** below verbatim (rendered as markdown).
+2. Stop. Do not perform any file reads, writes, or other tool calls.
+3. Do not generate any reports under `.wb/workflows/reports/`.
+
+Otherwise, ignore this section and proceed to the rest of the template.
+
+### HELP BLOCK — `/wbGit`
+
+## Three forms
+
+```
+/wbGit                             # draft commit message, no execution
+/wbGit --from-plan                 # infer commit from today's plan
+/wbGit --diff-file="diff.txt"      # read user-provided patch
+/wbGit --scan-recent               # infer from recently modified files
+/wbGit --execute                   # add + commit (asks for confirmation)
+/wbGit --push                      # push current branch (asks for confirmation)
+```
+
+Or use natural language: *"commit it"*, *"push"*, *"commit and push"*. Same result, AI parses intent.
+
+## When to run
+
+- After every logical unit of work (one task done, one bug fixed, one example added).
+- Before lunch / end of day — don't leave uncommitted work overnight.
+- After `/wbRelease` — to commit version bumps and tag the release.
+- After `/wbDeploy` — to commit any deploy-config changes.
+
+## When *not* to run
+
+- Mid-task, with half-done changes you'll throw away.
+- Just to "save progress" — that's what stash is for, not commit.
+- When tests are failing (commit broken code only if you're explicitly leaving a known-broken marker for tomorrow-you).
+
+## The good rhythm
+
+3-6 commits per working day. Less = you're batching too much (commits will mix purposes). More = you're commit-spamming (granular noise).
+
+## The mandatory provenance header
+
+Every Claude-generated commit message includes a `<!-- wbGit ... -->` block with five fields after the subject line: `date`, `model`, `session`, `plan`, `notes`. The block is HTML-comment-fenced so GitHub hides it visually but `git log` and `grep` still see it.
+
+Example (real format):
+
+```text
+fix(wb-dataviewer): apiResponse_ cache stale on route change
+
+<!-- wbGit
+date:    2026-04-25 11:08
+model:   claude-opus-4-7
+session: claude
+plan:    —
+notes:   resolves debug 20260419/debugs/apiResponse_loop.md
+-->
+
+Watcher key was reading apiResponse_ and writing it back...
+```
+
+Why every commit needs it: 6 months from now, `git blame` shows you which model drafted the message, which plan it advanced, and any salient notes (like "user authorized mixed-change override"). Without the header, you lose all that context.
+
+Full header schema lives in [wbGit_template_claude](../wbGit_template_claude.md). The `notes` field is the most useful day-to-day — fill it with anything future-you should know, otherwise write `—`.
+
+## What `/wbGit` will refuse to do
+
+- **Force push.** No `--force`. If your branch diverged, resolve it; don't overwrite.
+- **Skip hooks.** No `--no-verify`. If a hook fails, the hook is right or the code is right; either way, fix the underlying issue.
+- **Bundle unrelated changes silently.** If the diff has multiple unrelated purposes, the command surfaces this and asks you to split (or explicitly acknowledge the mix with a `chore:` message).
+- **Auto-push to main on a protected branch.** If main is protected and you don't have permission, the command surfaces the failure rather than attempting workarounds.
+
+## The mixed-change check
+
+If the AI sees that one diff spans:
+- A feature change
+- A bug fix
+- A docs update
+
+…it will refuse to commit as one and propose splitting. This is hygiene, not paranoia. Mixed commits are unreviewable in PR; conventional changelogs depend on per-commit categorization; `git revert` becomes destructive.
+
+If you genuinely want a mixed commit (rare — usually after a long forgotten branch), tell the AI explicitly: *"commit as mixed change"*. It will use a `chore:` prefix that acknowledges the mix.
+
+## The `/wbGit` ↔ closed-loop interaction
+
+`/wbGit` reads recent `reports/` to enrich commit messages. If a commit corresponds to a plan task, the message references the plan file. If it corresponds to a debug report, the message references the debug. This makes git history *queryable* against the report system.
+
+## When /wbGit is the wrong command
+
+- Stash unfinished work → use `git stash` directly (this is one of the few git commands you might do yourself; ask the user, don't reach for /wbGit).
+- Resolve a merge conflict → don't auto-resolve; do it manually.
+- Cherry-pick / rebase / interactive history rewrites → don't use `/wbGit`. These are user-driven operations.
+- Initial repo setup → done once, manually.
+
+`/wbGit` is for the daily commit/push rhythm. Anything more involved is yours.
+
+> For deeper reading: [`docs_claude/commands/wbGit/wbGit_practical_claude.md`](../../docs/docs_claude/commands/wbGit/wbGit_practical_claude.md) (or the `_eli5_`, `_expert_`, `_examples_` siblings).
+
+<!-- FLAGS_TABLE_START -->
+## Flags & shortcuts
+
+Both forms are equivalent — pass either:
+
+| Long form | Shortcut |
+|---|---|
+| `--execute` | `-e` |
+| `--force` | `-F` |
+| `--no-verify` | `-n` |
+| `--push` | `-p` |
+| `--from-plan` | `-P` |
+| `--diff-file` | `-d` |
+| `--scan-recent` | `-r` |
+
+`-h` / `--help` / `--h` (any command) prints this help block instead of executing.
+<!-- FLAGS_TABLE_END -->
+<!-- HELP_GATE_END -->
+
+<!-- FLAG_NORMALIZE_START -->
+## Flag normalization (apply BEFORE parsing args)
+
+Before processing `$ARGUMENTS`, normalize these short-form flags to their long equivalents:
+
+- `-e` → `--execute`
+- `-F` → `--force`
+- `-n` → `--no-verify`
+- `-p` → `--push`
+- `-P` → `--from-plan`
+- `-d` → `--diff-file`
+- `-r` → `--scan-recent`
+
+The rest of this template documents only the long forms; the substitution above is the only place short forms are mentioned.
+<!-- FLAG_NORMALIZE_END -->
+
+
+
+**ROLE:** The Archivist
+**TARGET:** The current working directory repository.
+
+## ━━━ OBJECTIVE ━━━
+Your job is to manage version control safely and professionally. You must analyze the current file changes and generate standard Conventional Commits.
+
+## ━━━ PHASE 1: RECONNAISSANCE ━━━
+1. Determine the source of changes based on flags:
+   - **Default**: You are forbidden from running terminal git commands. Ask the user for a summary.
+   - **`--from-plan`**: Read the most recent `plan_*.md` and extract `✅ Done` tasks.
+   - **`--diff-file="<path>"`**: Read the specified patch file directly.
+   - **`--scan-recent`**: Find and read files modified in the last 2 hours.
+2. Cross-reference the changes with any recent `plans/` or `standups/` to understand the context.
+
+## ━━━ PHASE 2: COMMIT GENERATION ━━━
+1. Draft a strict Conventional Commit message (e.g., `feat(wb-core): add logging capabilities`, `fix(dataviewer): resolve infinite loop in pagination`).
+2. Include an optional body explaining *why* the change was made if it is complex.
+
+## ━━━ PHASE 3: EXECUTION (Optional) ━━━
+If instructed by the user, execute the `git add`, `git commit`, and `git push` sequence. If generating a Pull Request, write a comprehensive PR description detailing the architectural changes.

package/templates/commands/wbHelp/wbHelp_template.md

@@ -0,0 +1,101 @@
+# /wbHelp: Catalog & Per-Command Help
+
+`/wbHelp` is a META-command. It does NOT generate a report. It prints help text and stops.
+
+---
+
+## Mode dispatch
+
+Inspect `$ARGUMENTS` and pick exactly one mode:
+
+| Args | Mode | Behavior |
+|---|---|---|
+| empty (no args) | **Catalog mode** | Print the role-grouped table of all `/wb*` commands |
+| `<wbX>` (one token, starts with `wb`) | **Per-command mode** | Output the HELP_GATE block of `wbX_template.md` (identical to `/wbX -h`) |
+| `<wbX> --help` / `<wbX> -h` | **Per-command mode** | Same as above; redundant flags ignored |
+| anything else | **Catalog mode** + a one-line note: *"Unrecognized form. Showing catalog. Use `/wbHelp <wbX>` for one command."* |
+
+**Normalization first**: strip a leading `/` from the command name (`/wbHelp wbAudit` and `/wbHelp /wbAudit` both work). Case-sensitive (`wbAudit`, not `wbaudit`).
+
+---
+
+## CATALOG MODE — render the table
+
+1. Read `../wb_commands_reference.claude.json`.
+2. Group commands by **role family** using the first noun of each command's `role` field. Map each `role` value to one of four families:
+   - 🧠 **Strategy / Planning** — `Architect`, `General`, `Product Manager`, `Navigator`, `Scrum Master`
+   - ✅ **Validation / Critique** — `Critic`, `Inspector`, `Diagnostician`, `Red Team`, `Quizmaster`, `Triage Officer`, `Synchronizer`
+   - 🔨 **Worker / Execution** — `Surgeon`, `Modernizer`, `Licensor`, `Tier Architect`, `Translator`, `Scribe`, `Release Engineer`, `Shipper`
+   - 📋 **Mechanical / Reporting** — `Tester`, `Janitor`, `Marketer`, `Stenographer`
+   
+   If a role doesn't fit any family cleanly, drop it under **🔨 Worker** by default.
+
+3. Output exactly this structure:
+
+```markdown
+# /wb* Command Catalog
+
+> 30 template-driven commands. Pass `--help` (or `-h`) to any command for its manual, or `/wbHelp <wbX>`.
+
+## 🧠 Strategy / Planning
+
+| # | Command | Role | Flags | What it does |
+|---|---|---|---|---|
+| 03 | `/wbPlan` | The General — coordinates workers and validators. | `--resume` (`-r`), `--scope` (`-s`), `--task` (`-t`) | Creates a structured task plan with worker/validator assignments. |
+| ...
+
+## ✅ Validation / Critique
+
+| # | Command | Role | Flags | What it does |
+| ...
+
+## 🔨 Worker / Execution
+
+| # | Command | Role | Flags | What it does |
+| ...
+
+## 📋 Mechanical / Reporting
+
+| # | Command | Role | Flags | What it does |
+| ...
+
+---
+
+**Need details on one command?** Run `/wbHelp <command>` (or `/<command> -h`).
+**Model picks per command:** see [`flow.wb-ui.com/concepts/model_recommendations_claude.md`](../../../../../apps/wb-flow/flow.wb-ui.com/concepts/model_recommendations_claude.md).
+```
+
+4. The `#` column uses the catalog's natural order (1, 2, 3 …) — read the JSON keys in their order of appearance and number sequentially. Don't invent positional numbers.
+
+5. The `What it does` column uses the JSON's `description` field, trimmed to a single sentence.
+
+6. The `Flags` column uses the JSON's `flags` array (each entry is `{"long": "--xxx", "short": "-x"}`). Render as comma-separated `` `--long` (`-short`) `` pairs. If the array is missing or empty, render `—` (em-dash). All commands also accept `-h` / `--help`; do NOT list that in this column — it's universal and would just be visual noise.
+
+7. **Do not** include the JSON's `phase`, `belt`, `recommendedModel`, or `references` fields in the table. Those belong in per-command help, not the catalog.
+
+8. Stop after printing. No file writes. No reports.
+
+---
+
+## PER-COMMAND MODE — delegate to the target template
+
+1. Compute `target = <wbX>` (the cleaned arg).
+2. Verify the file exists: `../<target>/<target>_template.md`. If not, output: *"Unknown command: `/<target>`. Run `/wbHelp` for the full catalog."* and stop.
+3. Read **only** the content between `<!-- HELP_GATE_START -->` and `<!-- HELP_GATE_END -->` in that template.
+4. Output it verbatim. No file writes. No reports.
+
+This is functionally identical to running `/<target> -h`. Same content, different invocation path.
+
+---
+
+## What `/wbHelp` will refuse to do
+
+- Generate any `.wb/workflows/reports/` file. This is a print-and-stop command.
+- Resolve flags or args meaningful to other commands. `/wbHelp wbGit --execute` shows the wbGit help; the `--execute` is ignored.
+- Run more than one mode in a single invocation.
+
+## When `/wbHelp` is the wrong command
+
+- You want the model-pick recommendations → `model_recommendations_claude.md`, not here.
+- You want to actually execute a command → drop the `wbHelp` prefix, just run `/wbX <args>`.
+- You want the full prose docs → `docs_claude/commands/<wbX>/<wbX>_practical_claude.md`.

package/templates/commands/wbIdea/wbIdea_template.md

@@ -0,0 +1,337 @@
+# wbIdea Template v1.1 — The Ideas Pipeline
+
+
+<!-- HELP_GATE_START -->
+## Help intercept (handle FIRST — before any other action)
+
+**If `$ARGUMENTS` contains `--help`, `-h`, or `--h`** (case-insensitive, anywhere in the args), DO NOT execute the command's normal procedure. Instead:
+
+1. Output the **HELP BLOCK** below verbatim (rendered as markdown).
+2. Stop. Do not perform any file reads, writes, or other tool calls.
+3. Do not generate any reports under `.wb/workflows/reports/`.
+
+Otherwise, ignore this section and proceed to the rest of the template.
+
+### HELP BLOCK — `/wbIdea`
+
+## Where `/wbIdea` fits in the pipeline
+
+```
+wbVision ──► wbIdea ──► wbPlan ──► wbWork ──► wbValid
+ (dream)    (capture)  (commit)   (execute)  (verify)
+```
+
+**`/wbVision`** = dream features (free-form, use rarely).
+**`/wbIdea`** = capture and score a concrete idea (tabular, trackable — but NOT committed to execution).
+**`/wbPlan`** = commit to execution (worker/validator assignments, DAG dependencies).
+
+The key distinction: ideas have a **Score** and a **☐ Valid** column, but no worker/validator assignments. An idea becomes a task only when explicitly promoted via validation.
+
+## The four forms
+
+```
+/wbIdea <pkg>                               # AI scans context and proposes ideas
+/wbIdea <pkg> --task="<description>"         # Register a specific idea manually
+/wbIdea <pkg> --resume                       # Re-read existing idea file, re-score
+/wbIdea <pkg> --id=1 --promote               # Promote idea #1 to today's plan file
+/wbIdea <pkg> --id=1,2 --reject              # Mark ideas 1,2 as 🚫 Rejected
+/wbIdea <pkg> --id=1 --defer                 # Mark idea 1 as ⏸️ Deferred
+```
+
+## Reading an idea file
+
+Every idea file has a table. Four columns matter most:
+
+- **Score** — 1–10 advisory score. How strongly this idea should become a task. 10 = "implement immediately". 1 = "nice-to-have, maybe never".
+- **Idea** — must be specific enough that another session could evaluate it without re-reading the source.
+- **☐ Valid** — validation state. Determines whether the idea should be promoted, deferred, or rejected.
+- **→ Task** — the bridge to `/wbPlan`. When an idea is promoted, this column becomes a link to the plan row.
+
+Skim these before deciding. If the ideas are vague, the file is weak; re-run with a more specific scope.
+
+## When to run
+
+- Anytime you have a speculative improvement that isn't worth committing to immediately.
+- After `/wbVision` to formalize its best ideas into trackable rows.
+- After `/wbAudit` with low-severity findings that are improvements rather than bugs.
+- When you want to build a backlog of possibilities without creating plan overhead.
+
+## When NOT to run
+
+- The idea fits in one sentence and you're ready to execute → just describe the task.
+- You have an existing plan with open tasks → finish those first.
+- You want to brainstorm broadly → use `/wbVision`.
+
+## When /wbIdea is the wrong command
+
+- You want to *do* the work → describe the task directly, or `/wbWork`.
+- You want to *commit* to execution → `/wbPlan`.
+- You want to *brainstorm* with no structure → `/wbVision`.
+- You want to *audit* code quality → `/wbAudit`.
+
+`/wbIdea` answers one question: *"Is this worth doing eventually?"* Not "what are the steps?" (that's `/wbPlan`) and not "what could we build?" (that's `/wbVision`).
+
+> For deeper reading: [`docs_claude/commands/wbIdea/wbIdea_practical_claude.md`](../../docs/docs_claude/commands/wbIdea/wbIdea_practical_claude.md) (or the `_eli5_`, `_expert_`, `_examples_` siblings).
+
+<!-- FLAGS_TABLE_START -->
+## Flags & shortcuts
+
+Both forms are equivalent — pass either:
+
+| Long form | Shortcut | Description |
+|---|---|---|
+| `--resume` | `-r` | Re-read existing idea file, re-score |
+| `--scope` | `-s` | Scope override |
+| `--task` | `-t` | Explicit idea description |
+| `--id` | `-i` | Specifies idea indices to target |
+| `--promote` | `-p` | Promote idea(s) to today's plan file |
+| `--reject` | `-x` | Mark idea(s) as 🚫 Rejected |
+| `--defer` | `-d` | Mark idea(s) as ⏸️ Deferred |
+
+`-h` / `--help` / `--h` (any command) prints this help block instead of executing.
+## Self-correct mode (dual-mode invocation)
+
+```
+/wbIdea <scope_folder>           # normal mode — produce a fresh output file
+/wbIdea <previous_output_file>   # self-correct mode — verify & repair the file in place
+```
+
+When the first arg is an existing output file from a prior `/wbIdea` run (detected by its first H1 — see this template's **Detection** section), the command runs in **verify-and-repair** mode: gap-fills missing fields, normalizes links, ticks done/valid checkboxes whose reports exist, never rewrites authored content. See [`../_shared/output_conventions.md`](../_shared/output_conventions.md) §3.
+
+<!-- FLAGS_TABLE_END -->
+<!-- HELP_GATE_END -->
+
+<!-- FLAG_NORMALIZE_START -->
+## Flag normalization (apply BEFORE parsing args)
+
+Before processing `$ARGUMENTS`, normalize these short-form flags to their long equivalents:
+
+- `-r` → `--resume`
+- `-s` → `--scope`
+- `-t` → `--task`
+- `-i` → `--id`
+- `-p` → `--promote`
+- `-x` → `--reject`
+- `-d` → `--defer`
+
+The rest of this template documents only the long forms; the substitution above is the only place short forms are mentioned.
+<!-- FLAG_NORMALIZE_END -->
+
+
+
+> **How to use**: Copy the prompt below, replace `__PLACEHOLDERS__`, paste to any AI agent.
+> Everyone gets the SAME prompt. Each agent finds their alias and does only their tasks.
+> **Read first:** [`../_shared/output_conventions.md`](../_shared/output_conventions.md) — applies to every cell of the output, including **§9 Action Type Tagging**: declare `type:` + `emits:` in YAML front-matter, add a plain-text `Requires` column to every ideas/candidates table (each idea gets a 🧠/✅/🔨/📋 tag indicating what role would carry it out), include a `## 🔗 Action Types` legend.
+
+---
+
+## Filename & Folder Convention
+
+**Path:** `<target_folder>/.wb/workflows/reports/<YYYY>/<MM>/<DD>/ideas/idea_<target>_<YYYYMMDD>.md`
+
+**STRICT NAMING RULE:** The idea filename MUST be exactly `idea_<folder_scope>_<YYYYMMDD>.md`. Do NOT append source types, issue names, or any other suffixes.
+
+**ONE FILE PER DAY rule:** There must be exactly ONE idea file per day per folder scope. All ideas — regardless of origin (audit, vision, manual, etc.) — go into the same file as separate Entry #N sections or rows in the same idea table.
+
+**MERGE-OR-CREATE protocol:**
+1. If `idea_<folder>_<YYYYMMDD>.md` already exists → APPEND your ideas as a new Entry #N section (tagged `*(ModelName — HH:MM)*`).
+2. If it does not exist → CREATE it.
+3. If multiple `idea_*<YYYYMMDD>*.md` files exist for the same day (e.g., from a prior bug) → MERGE them into one `idea_<folder>_<YYYYMMDD>.md`, combining all idea tables.
+
+---
+
+## Detection (Self-Correct Mode)
+
+Trigger self-correct when the input file's first H1 matches:
+`# Idea Backlog: <scope> — <YYYY-MM-DD>`
+
+Behavior is defined in [`../_shared/output_conventions.md`](../_shared/output_conventions.md) §3.
+Idea-specific gap-fills:
+- Blank `☐ Done` for an idea whose `ideas_reports/idea_<N>/idea_<N>_report_*.md` exists → check it (`✅<br><worker>`).
+- Blank `☐ Valid` for an idea whose exploration report has a validator score appended → fill `✅ <Score>/10<br><validator>`.
+- Plain-text file references → convert to relative markdown links per §1.
+- Missing `Suggested By` for ideas with known origin commands → fill from context.
+- **Column completeness — verify ALL required columns are present** (see Column Rules above): `#`, `Requires`, `Score`, `🔗`, `Idea`, `P`, `Est. Time (mins)`, `Suggested By`, `☐ Done`, `☐ Valid`, `→ Task`. If any column is missing — most commonly **`Requires`** in idea files written before v1.1 — insert it after `#` and back-fill every row by inferring the tag from the idea's wording.
+- **Legend section check:** if `## 🔗 Action Types` legend is absent, append it just before the Generated Files footer (template in `_shared/output_conventions.md` §9.3).
+- **Link beautification (per `_shared/output_conventions.md` §1.1):** walk EVERY markdown link in the file (front-matter, callouts, prose, table cells, footer). Apply the four-rule detection: rewrite any link where (1) label == href, (2) label contains a `/` other than a single trailing slash, (3) label has `..`/`…`/`./`, or (4) label is absolute. Keep the href intact; replace the label with the basename of the href. Example: `[../../plans/plan_X.md](../../plans/plan_X.md)` → `[plan_X.md](../../plans/plan_X.md)`.
+- **Path correctness (per `_shared/output_conventions.md` §1.2):** for every link to a report file or scope-root file, derive the href from §1.2's canonical-path table — do NOT count `../` segments by hand. The most common break is prev-day links across a month boundary (need `../../../<MM-prev>/<DD-prev>/...`, not `../../<DD-prev>/...`). Rewrite mismatches.
+
+---
+
+## 🚀 The Ideation Prompt (copy from here ↓)
+
+```
+━━━━━━━━━━━━━ /wbIdea v1.1 ━━━━━━━━━━━━━
+
+📁 PROJECT: __PROJECT_NAME__
+📅 DATE: __TODAY__
+🤖 MODEL: __YOUR_MODEL_NAME__
+🖥️ CLIENT: __YOUR_CLIENT__
+📂 IDEA FILE: __TARGET_FOLDER__/.wb/workflows/reports/__YYYY__/__MM__/__DD__/ideas/idea___TARGET_NAME_____YYYYMMDD__.md
+
+━━━ CONTEXT ━━━
+
+Read the `.wb/workflows/context.md` file in the target folder(s) FIRST.
+Check `../model_recommendations.md` for role assignments.
+Read `../_shared/output_conventions.md` — your output MUST follow it.
+
+━━━ INPUT MODE DETECTION ━━━
+
+If the input is a file matching the idea-output schema (first H1 = "# Idea Backlog: <scope> — <date>"):
+  - If target IDs and state flags (`--promote`, `--reject`, `--defer`) are provided:
+    → **STATE OVERRIDE MODE**: Directly edit the idea file.
+      - `--promote`: Set `☐ Valid` to `🎯 Promoted` and trigger the Promotion Protocol (see below).
+      - `--reject`: Set `☐ Valid` to `🚫 Rejected`.
+      - `--defer`: Set `☐ Valid` to `⏸️ Deferred`.
+    Do NOT generate new ideas.
+  - Else:
+    → **SELF-CORRECT MODE** (see _shared/output_conventions.md §3).
+Else:
+  → **FRESH-IDEA MODE** (proceed with the prompt below).
+
+━━━ THE CONTEXT ━━━
+
+"""
+__YOUR_MESSY_DESCRIPTION_OR_EMPTY__
+"""
+
+━━━ IDEA TABLE (v1.1 — Scored Ideas Pipeline, Action-Tagged) ━━━
+
+| # | Requires | Score | 🔗 | Idea | P | Est. Time (mins) | Suggested By | ☐ Done | ☐ Valid | → Task |
+|---|---|---|---|---|---|---|---|---|---|---|
+| 1 | 🔨 Worker | 8 | <span title="Run: /wbExplain idea_<scope>_<date>.md --id=1 --scope=idea">📄</span> | [idea description] | P1 | 30 | <ModelName> via /wbIdea | ⬜ | ⬜ | — |
+| 2 | 🧠 Planner | 5 | <span title="Run: /wbExplain idea_<scope>_<date>.md --id=2 --scope=idea">📄</span> | [idea description] | P2 | 15 | <ModelName> via /wbIdea | ⬜ | ⬜ | — |
+
+**Column Rules:**
+1. **# (Idea Index)**: When an idea is explored (Done = ✅), this index MUST become a markdown link to the exploration report (e.g. `[1](ideas_reports/idea_1/idea_1_report_<scope>_<YYYYMMDD>.md)`).
+2. **Requires (Action Type)**: Plain-text tag indicating which canonical role would carry the idea out *if it were promoted to a task* — one of `🧠 Planner` / `✅ Validator` / `🔨 Worker` / `📋 Mechanical`. Per `_shared/output_conventions.md` §9.3 this column is mandatory for every idea file; the file MUST also include a `## 🔗 Action Types` legend section just before the Generated Files footer. Tag selection rule: pick the role the *promoted task* would require — a code edit / refactor → 🔨 Worker; a "should we?" / multi-step decomposition → 🧠 Planner; a quality / safety review → ✅ Validator; a "run command and capture …" → 📋 Mechanical. Hybrid ideas MUST be split into two rows.
+3. **Score (Advisory Score)**: Integer 1–10. Computed by the suggesting model based on: **impact × feasibility × urgency**. 10 = "implement immediately". 1 = "nice-to-have, maybe never". The score informs validation but does not auto-promote.
+4. **🔗 (Details)**: Contextual Lazy Loading trigger. Inactive: `<span title="Run: /wbExplain idea_<scope>_<date>.md --id=<N> --scope=idea">📄</span>`. If an explanation file has been generated, upgrade to an active link: `[📄](ideas_reports/idea_<N>/idea_<N>_details_<scope>_<YYYYMMDD>.md "View Details")`.
+5. **Idea**: Short imperative description. Clear enough to be promoted to a plan task without re-reading the source.
+6. **P (Priority)**: P0–P3. Indicates the *urgency of evaluating* the idea, not of executing it.
+7. **Est. Time (mins)**: How long the *task* would take if this idea were promoted. Not the time to evaluate the idea.
+8. **Suggested By**: Who/what proposed this idea. Format: `<ModelName> via <command>` (e.g., `Opus 4.7 via /wbAudit`, `Manual`, `Gemini 3.1 Pro via /wbVision`).
+9. **☐ Done (Exploration Status)**: `⬜` = not explored. `✅<br>ModelName` = idea has been explored and documented via `/wbWork idea_*.md --id=N`. Unlike plan Done, this means "explored", not "implemented".
+10. **☐ Valid (Validation Status)**: The verdict scale:
+   - `⬜` = not reviewed
+   - `🎯 Promoted <Score>/10<br>ModelName` = "This deserves to be a task" (Score 8–10). Triggers promotion protocol.
+   - `✅ <Score>/10<br>ModelName` = "Good idea, worth doing when capacity allows" (Score 5–7). No auto-promotion.
+   - `⏸️ Deferred <Score>/10<br>ModelName` = "Interesting but not the right time" (Score 3–4).
+   - `🚫 Rejected <Score>/10<br>ModelName` = "Not worth pursuing" (Score 1–2).
+   Validation is cumulative (same as plan Valid): append new validations below existing ones.
+11. **→ Task (Promotion Link)**: `—` when not promoted. When promoted, becomes a relative markdown link: `[→ Plan #M](../plans/plan_<scope>_<YYYYMMDD>.md)`.
+
+━━━ ENTRY HEADER FORMAT ━━━
+
+Each idea batch added to the file MUST have a section header:
+
+## 💡 Ideas — <origin_description> *(<ModelName> via <Client> — <HH:MM>)*
+> **Source:** <relative link to source file if applicable>
+> **Model:** <model_name>
+> **Date:** <YYYY-MM-DD>
+
+Example origins:
+- "AI-generated ideas for wb-dataviewer"
+- "Low-severity audit findings from /wbAudit"
+- "Feature proposals from /wbVision"
+- "Manual idea registration"
+
+━━━ SCORING HEURISTIC ━━━
+
+When computing the Score (1–10), use this matrix:
+
+| Factor | Weight | Anchors |
+|---|---|---|
+| **Impact** | 40% | 10 = fundamentally changes UX/DX. 5 = noticeable improvement. 1 = cosmetic. |
+| **Feasibility** | 30% | 10 = straightforward, low risk. 5 = moderate complexity. 1 = requires architectural change. |
+| **Urgency** | 30% | 10 = blocking other work. 5 = would be nice this sprint. 1 = no timeline pressure. |
+
+`Score = round(impact × 0.4 + feasibility × 0.3 + urgency × 0.3)`
+
+━━━ PROMOTION PROTOCOL ━━━
+
+When an idea is promoted (via `--promote` flag or via `/wbValid idea_*.md --id=N` with `🎯 Promoted` verdict):
+
+1. Locate (or create) today's plan file: `<target>/.wb/workflows/reports/<YYYY>/<MM>/<DD>/plans/plan_<scope>_<YYYYMMDD>.md`.
+2. Append the idea as a new task row:
+   - `Origin` = `💡 [/wbIdea #<N>](../ideas/idea_<scope>_<YYYYMMDD>.md)` (relative link).
+   - `Task` = copy from the `Idea` column.
+   - `P` = copy from the idea's `P` column.
+   - `Est. Time` = copy from the idea's `Est. Time` column.
+   - `Worker` / `Validator` = fill from `model_recommendations.md`.
+   - `☐ Done` = `⬜`.
+   - `☐ Valid` = `⬜`.
+3. Update the idea file:
+   - `→ Task` column = `[→ Plan #M](../plans/plan_<scope>_<YYYYMMDD>.md)`.
+   - `☐ Valid` = `🎯 Promoted <Score>/10<br>ModelName` (if not already set).
+
+━━━ SOURCE LINKING ━━━
+
+When ideas originate from another command's output (e.g. an audit, a vision), the entry header MUST include a relative markdown link to that source file. Example:
+
+> ## 💡 Ideas — /wbVision findings *(Gemini 3.1 Pro — 14:30)*
+> - **Source:** [vision_wb-core_20260508.md](../visions/vision_wb-core_20260508.md) Entry #1
+> - **Origin Command:** `/wbVision packages/wb-core/`
+> - **Ideas registered:** 3
+
+━━━ "WHAT NEXT" ━━━
+
+In **Fresh-Idea Mode**, end the idea file with:
+
+> ## 🧭 What's Next?
+> - Run [`/wbWork idea_<scope>_<YYYYMMDD>.md --id=<N>`] to explore idea #N in depth.
+> - Run [`/wbExplain idea_<scope>_<YYYYMMDD>.md --id=<N> --scope=idea`] for a deep-dive explanation.
+> - Run [`/wbValid idea_<scope>_<YYYYMMDD>.md --id=<N>`] to validate and potentially promote.
+> - Run [`/wbIdea <target> --id=<N> --promote`] to directly promote a high-scoring idea.
+
+In **Self-Correct Mode**, if you are re-checking an existing idea file:
+1.  **Analyze current progress**: Check which ideas are explored, validated, promoted.
+2.  **Re-score ideas** if context has changed.
+3.  **Suggest immediate next actions** based on the current state.
+
+━━━ AUTO-APPEND FOOTER ━━━
+
+At the VERY END of the file (after "What's Next?"), you MUST append the `## 📂 Generated Files (__YYYYMMDD__)` cross-link footer. Do NOT use simple tables. You MUST use the rich "Tier 1" layout from `_shared/output_conventions.md` §5.
+
+Format required:
+```markdown
+---
+## 📂 Generated Files (__YYYYMMDD__)
+> Auto-appended per `_shared/output_conventions.md` §5. Same-level snapshot of top-level command outputs at write time.
+
+### 📚 Base Reference Files
+| Type | File | Description |
+|---|---|---|
+| Foundational | [context.md](../../../../../context.md) | Permanent Identity and Architecture (Source of Truth) |
+| Snapshot | [context_<scope>_<date>.md](../contexts/context_<scope>_<date>.md) | Daily snapshot used for current session context |
+| Foundational | [dev.md](../../../../../dev.md) | Permanent Development Commands and Status |
+
+### Local Files
+
+| Category | File (day N) | File (day N-1) | Source Command |
+|---|---|---|---|
+| Identity | [context.md](../../../../../context.md) | — | Foundational Identity & Architecture |
+| Identity | [dev.md](../../../../../dev.md) | — | Foundational Dev Commands & Status |
+| Reports | [audit_<scope>_<date>.md](../audits/audit_<scope>_<date>.md) | [audit_<scope>_<prev-date>.md](../../<prev-DD>/audits/audit_<scope>_<prev-date>.md) | `/wbAudit` |
+| Reports | [plan_<scope>_<date>.md](../plans/plan_<scope>_<date>.md) | [plan_<scope>_<prev-date>.md](../../<prev-DD>/plans/plan_<scope>_<prev-date>.md) | `/wbPlan` |
+| Reports | **idea_<scope>_<date>.md** *(this file)* | [idea_<scope>_<prev-date>.md](../../<prev-DD>/ideas/idea_<scope>_<prev-date>.md) | `/wbIdea` |
+| Reports | [next_<scope>_<date>.md](../nexts/next_<scope>_<date>.md) | [next_<scope>_<prev-date>.md](../../<prev-DD>/nexts/next_<scope>_<prev-date>.md) | `/wbNext` |
+
+### Global Files (`core2/` monorepo root)
+
+| Category | File (day N) | File (day N-1) | Source Command |
+|---|---|---|---|
+| Reports | [audit_core2_<date>.md](../../../../../../../../../../.wb/workflows/reports/<YYYY>/<MM>/<DD>/audits/audit_core2_<date>.md) | [audit_core2_<prev-date>.md](../../../../../../../../../../.wb/workflows/reports/<YYYY>/<MM>/<prev-DD>/audits/audit_core2_<prev-date>.md) | `/wbAudit core2/` |
+| Reports | [plan_core2_<date>.md](../../../../../../../../../../.wb/workflows/reports/<YYYY>/<MM>/<DD>/plans/plan_core2_<date>.md) | [plan_core2_<prev-date>.md](../../../../../../../../../../.wb/workflows/reports/<YYYY>/<MM>/<prev-DD>/plans/plan_core2_<prev-date>.md) | `/wbPlan core2/` |
+| Tracks | [track_core2_<date>.md](../../../../../../../../../../.wb/workflows/tracks/<YYYY>/<MM>/<DD>/track_core2_<date>.md) | [track_core2_<prev-date>.md](../../../../../../../../../../.wb/workflows/tracks/<YYYY>/<MM>/<prev-DD>/track_core2_<prev-date>.md) | `/wbTrack core2/` |
+
+<details>
+  <summary>📂 Sub-Package: [Active Package Name]</summary>
+
+| Category | File (day N) | File (day N-1) | Source Command |
+|---|---|---|---|
+| Reports | [audit_subpkg_<date>.md](../../../../../../../../../../apps/wb-core/subpkg/.wb/workflows/reports/<YYYY>/<MM>/<DD>/audits/audit_subpkg_<date>.md) | [audit_subpkg_<prev-date>.md](../../../../../../../../../../apps/wb-core/subpkg/.wb/workflows/reports/<YYYY>/<MM>/<prev-DD>/audits/audit_subpkg_<prev-date>.md) | `/wbAudit` |
+
+</details>
+```
+```