wb-flow

Appearance

/wbIdea — Examples ​

/wbIdea captures speculative ideas with scores. This file shows three scenarios: registering ideas natively, ideas from /wbVision, and the full promote-to-plan lifecycle.

Example 1 — /wbIdea packages/wb-dataviewer ​

text
$ /wbIdea packages/wb-dataviewer

[AI] Loading context.md...
[AI] Loading recent reports/...
[AI] - Last audit: clean, 2 P3 cosmetic findings
[AI] - Active plan: 3/5 tasks done
[AI]
[AI] Writing reports/2026/05/08/ideas/idea_wb-dataviewer_20260508.md:
markdown
# Idea Backlog: wb-dataviewer — 2026-05-08

## 💡 Ideas — AI-generated for wb-dataviewer *(the AI agent via Antigravity — 14:30)*
> **Model:** the AI agent
> **Date:** 2026-05-08

| # | Score | 🔗 | Idea | P | Est. Time (mins) | Suggested By | ☐ Done | ☐ Valid | → Task |
|---|---|---|---|---|---|---|---|---|---|
| 1 | 8 | 📄 | Add column-level search with multi-color highlighting | P1 | 45 | the AI agent via /wbIdea | ⬜ | ⬜ | — |
| 2 | 6 | 📄 | Support row grouping by any column value | P2 | 120 | the AI agent via /wbIdea | ⬜ | ⬜ | — |
| 3 | 4 | 📄 | Add CSV/XLSX export button to toolbar | P2 | 30 | the AI agent via /wbIdea | ⬜ | ⬜ | — |
| 4 | 9 | 📄 | Implement virtual scrolling for 10k+ row datasets | P1 | 240 | the AI agent via /wbIdea | ⬜ | ⬜ | — |

What's worth noticing:

  • Scores range from 4 to 9. The model didn't inflate everything to 8+.
  • Idea #4 (virtual scrolling) scores highest because it's high-impact + high-urgency for large datasets.
  • Idea #3 (CSV export) scores 4 — nice-to-have but low urgency.

Example 2 — Ideas auto-registered from /wbVision ​

text
$ /wbVision packages/wb-core

[AI] Writing vision file...
[AI] Auto-registering 3 ideas in idea_wb-core_20260508.md...

The idea file now has a /wbVision-sourced entry:

markdown
## 💡 Ideas — /wbVision proposals *(the AI agent via Cline — 10:15)*
> **Source:** `vision_wb-core_20260508.md`
> **Origin Command:** `/wbVision packages/wb-core/`
> **Ideas registered:** 3

| # | Score | 🔗 | Idea | P | Est. Time (mins) | Suggested By | ☐ Done | ☐ Valid | → Task |
|---|---|---|---|---|---|---|---|---|---|
| 1 | 7 | 📄 | Server-driven view config (admin pushes layout without rebuild) | P2 | 120 | the AI agent via /wbVision | ⬜ | ⬜ | — |
| 2 | 9 | 📄 | Diff view (compare two dataset snapshots row-by-row) | P1 | 480 | the AI agent via /wbVision | ⬜ | ⬜ | — |
| 3 | 3 | 📄 | Dark mode theme tokens | P3 | 30 | the AI agent via /wbVision | ⬜ | ⬜ | — |

What's worth noticing:

  • Each idea links back to the vision file.
  • Scores reflect the vision's own Effort/Value assessment (diff view = high value, scored 9).

Example 3 — Full lifecycle: Explore → Validate → Promote ​

Step 1: Explore idea #4 ​

text
$ /wbWork idea_wb-dataviewer_20260508.md --id=4

[AI] Reading idea #4: "Implement virtual scrolling for 10k+ row datasets"
[AI] Exploring...
[AI] Writing ideas_reports/idea_4/idea_4_report_wb-dataviewer_20260508.md
[AI] Updated ☐ Done to ✅

Step 2: Validate idea #4 ​

text
$ /wbValid idea_wb-dataviewer_20260508.md --id=4

[AI] Reading exploration report...
[AI] Verdict: 🎯 Promote (9/10)
[AI] "Virtual scrolling is critical for the 10k+ dataset use case.
[AI] Feasibility is high — existing row rendering can be wrapped.
[AI] Recommend P1 priority in plan."
[AI] Appended validation to idea_4_report.
[AI] Updated ☐ Valid to: 🎯 Promoted 9/10
[AI] Executing Promotion Protocol...
[AI] → Added as Task #6 in plan_wb-dataviewer_20260508.md
[AI] → Updated → Task column: `→ Plan #6`

The idea file now shows: ​

markdown
| `4` | 9 | 📄 | Virtual scrolling for 10k+ rows | P1 | 240 | the AI agent via /wbIdea | ✅<br>the AI agent | 🎯 Promoted 9/10<br>the agent 4.6 | `→ Plan #6` |

The plan file now shows (auto-ingested): ​

markdown
| 6 | 💡 `/wbIdea #4` | Virtual scrolling for 10k+ rows | /wbTest packages/wb-dataviewer --scope=task-6 | P1 | 240 | the AI agent · ~$0.20 | the AI agent | ⬜ | ⬜ |

The pattern ​

  1. Register → ideas appear with scores in idea_*.md.
  2. Explore → /wbWork idea_*.md --id=N writes a feasibility report.
  3. Validate → /wbValid idea_*.md --id=N assigns a verdict (promote/recommend/defer/reject).
  4. Promote → 🎯 Promoted ideas auto-ingested by /wbPlan as task rows.
  5. Execute → normal /wbWork plan_*.md --id=N on the promoted task.