mde-linear
暂无简介
- License 许可
- MIT
- 所属仓库
- amoai-tech/mdeai
- 最近更新
- 2026-11-13
- Stars
- ★ 2
- Forks
- 1
02 资源文件 2
- 参考文档evals/evals.json1.1 KB
- 参考文档references/mdeai-overrides.md2.7 KB
03 SKILL.md 正文
⚠️ DEPRECATED (2026-06-20) — superseded by the canonical
linearskill. All content frommde-linearwas merged into.agents/skills/linear/(SKILL.md + references/). Use thelinearskill instead. Kept temporarily for review; scheduled for removal once the migration is approved. Do not extend this file.
mde-linear — mdeai Linear operations
BLUF: Load linear.md first, then this skill for mdeai-specific rules. For API/CLI mechanics, delegate to the generic linear skill.
Issue descriptions are execution contracts — not title-only stubs. Every executable SAN-### gets completion steps + Gantt (and sequence diagram when HITL/routing).
When to invoke
| Trigger | Action |
|---|---|
| "Create Linear issue for …" | Follow creation workflow + full description skeleton below |
| "Update SAN-###" / "Mark SAN-### Done" | Pair ID with title; verify gates; sync checkboxes |
| "Add steps / Gantt to SAN-###" | Load linear-issue-steps.md |
| Cycle / milestone / initiative update | Follow linear.md §Cycles |
| "Sync tasks to Linear" | Bulk script or MCP per linear.md |
| CK-V2 / VEB / program tasks | Add skills table + link to program tracker |
mdeai rules (always)
- Never bare
SAN-NNN— alwaysSAN-NNN · <full Linear title>in chat, commits, PRs, docs. - Branch naming:
ai/san-NNN-spec-id-slug(see linear.md). - PR magic words:
Closes SAN-NNNin PR body. - Deprecated prefixes: Do not create new
SCREEN-*,EVP-*,IMP-*issues. - Phase labels:
phase:launch(Cycle 1 P0),phase:mvp(post-launch). - One worktree, one PR per SAN issue (app code).
Executable issue description (required skeleton)
Paste into save_issue description when creating or refreshing an issue:
## <SPEC-ID> — <short title>
**In plain terms:** <Who notices what on which surface — one sentence>
**Blocked by:** SAN-### · <full title> (or "—")
**Skills (.claude/skills — load ≤5):**
| Skill | Why (persona-visible) |
|---|---|
| `mde-task-lifecycle` | Five-phase gates + ship bookkeeping |
| … | … |
Map: `<path to program skills map if exists>`
---
### Completion steps (check in order)
#### A. <Gate / scope>
- [ ] **A1** <action> — proof: <file, grep, URL, or behavior>
#### B. <Build>
- [ ] **B1** …
#### E. Verify + ship
- [ ] **E1** Vitest / unit tests
- [ ] **E2** Playwright or localhost proof
- [ ] **E3** `npm run floor` green
- [ ] **E4** Evidence `docs/tasks/testing/evidence/SAN-###/`
- [ ] **E5** PR merged · `Closes SAN-###` · Linear Done
---
### Gantt — SAN-### step sequence
```mermaid
gantt
title SAN-### · SPEC-ID — persona-visible title
dateFormat YYYY-MM-DD
axisFormat %b %d
section Gate
Blocker cleared :done, s0, 2026-06-12, 1d
section Build
Core deliverable :active, s1, after s0, 3d
section Verify
floor + merge :crit, s2, after s1, 2d
Done :milestone, sdone, after s2, 0d
Unblocks: SAN-### · <next task>
**Canonical reference:** [mde-task-lifecycle/references/linear-issue-steps.md](../mde-task-lifecycle/references/linear-issue-steps.md)
---
## Plain-language descriptions (persona-first)
Write so a busy non-engineer knows what changes in the product.
| Bad (jargon) | Good (mdeai) |
|---|---|
| "Migrate `useCoAgent` → `useAgent`" | "**Roberto**'s event wizard keeps draft state when the agent fills fields" |
| "Flag-gated v2 provider" | "Prod stays v1 until env flag on — **Camila**'s `/chat` untouched this PR" |
| "Wire `useRenderTool`" | "Sales KPI cards still read tool **result**, not LLM prose" |
**Personas:** Roberto `/host/*` · Camila `/chat` `/rentals` · Andrés checkout · Patricia `/admin` · Sofía CI · Lucía Playwright.
Each checklist step's proof should name who breaks if the step is skipped.
---
## Mermaid diagrams in Linear
Load **`mermaid-diagrams`** before authoring. Validate against `references/gantt.md` and `references/best-practices.md`.
| Diagram | When | Rules |
|---|---|---|
| **Gantt** | Every executable issue | `dateFormat YYYY-MM-DD` · sections = checklist phases · `:crit` = merge/floor · `:milestone` = Done |
| **flowchart** | Program deps (epic) | `accTitle` + `accDescr` for a11y |
| **sequenceDiagram** | HITL, routing, webhooks | Agent ↔ UI ↔ user; e.g. Roberto publish approval |
**Program Gantt** (second diagram on cutover/epic issues): chain SAN-888 → 889 → 890 → 891 with parallel SAN-892.
**CK-V2 examples:** Linear SAN-888–892 (verified 2026-06-12).
### Gantt status tokens
| Token | Meaning |
|---|---|
| `:done` | Shipped / merged |
| `:active` | Current branch / PR |
| `:crit` | Blocks Done (floor, HITL, highest risk) |
| `:milestone` | Linear Done (0d) |
| `after <id>` | Sequential dependency within issue |
---
## Progress tracker sync (programs)
When a program has repo trackers, keep **Linear + docs** aligned:
| Artifact | Typical path | Update when |
|---|---|---|
| Linear checkboxes | Issue description | Step completes |
| Progress tracker | `docs/<program>/todo.md` | 🟢🟡🔴⚫ + % per task |
| Changelog | `docs/<program>/changelog.md` | Merge, proof, blocker |
**CK-V2:** [docs/upgradeV2/todo.md](../../../docs/upgradeV2/todo.md) · [changelog.md](../../../docs/upgradeV2/changelog.md)
Run **`task-verifier`** before marking steps `[x]` and setting state **Done**.
---
## Issue creation workflow
1. Read [linear.md](../../../linear.md) for project, labels, cycle.
2. Resolve full title (grep existing issues — no duplicate spec IDs).
3. Draft description with skeleton above (steps + Gantt + skills).
4. Set `team`, `project`, labels (`phase:launch` if P0), estimate if known.
5. Create via MCP `save_issue` or `linear issue create`.
6. Reply with linked title: `[SAN-NNN · Title](https://linear.app/sanjiovani/issue/SAN-NNN/...)`.
7. If program task: add row to `docs/<program>/todo.md`.
---
## Issue update / Done workflow
[ ] All completion steps E1–E5 checked (or N/A documented) [ ] task-verifier pass on disk claims [ ] Localhost proof (gate 9) unless pure-doc task [ ] PR merged with Closes SAN-NNN [ ] todo.md dot 🟢 + changelog dated entry [ ] Linear state → Done [ ] Gantt: mark :done / :milestone on final items (optional comment)
**Do not** flip Done on floor fail, missing evidence, or open blocker — use 🟡/🔴 in tracker instead.
---
## CK-V2 program quick map
| SAN | SPEC | Persona / surface | Flag |
|---|---|---|---|
| 886 | CK-V2-000 | Epic — no code | — |
| 887 | CK-V2-001 | Sofía — spike doc | — |
| 888 | CK-V2-002 | Roberto `/host/analytics` | `COPILOTKIT_V2_ANALYTICS` |
| 889 | CK-V2-003 | Roberto `/host/event/new` | `COPILOTKIT_V2_HOST_EVENT` |
| 890 | CK-V2-004 | Camila `/chat` | `COPILOTKIT_V2_CHAT` |
| 891 | CK-V2-005 | All — retire react-ui | remove flags |
| 892 | CK-V2-006 | Patricia — tag build-on-v2 | — |
| 893 | CK-V1-001 | Roberto `/host/event/new` (investigate) | — |
| 894 | CK-V1-002 | Roberto `/host/event/new` (fix max-depth) | — |
| 895 | CK-V2-007 / CK-AI-001 | Roberto — `thought_signature` hygiene | — |
| 896 | CK-V2-008 | Sofía — evidence refresh | — |
| 897 | CK-V2-009 | Roberto preview `/host/analytics` | `COPILOTKIT_V2_ANALYTICS` preview only |
| 898 | CK-V2-010 | Roberto — v2 hydration fix | `COPILOTKIT_V2_HOST_EVENT` |
| 899 | CK-AI-002 | All agents — memory signature round-trip | — |
Skills map: [docs/upgradeV2/CK-V2-skills-map.md](../../../docs/upgradeV2/CK-V2-skills-map.md)
Lifecycle: [copilotkit-v2-migration.md](../mde-task-lifecycle/references/copilotkit-v2-migration.md)
---
## MCP / CLI delegation
For raw API operations, load the generic **`linear`** skill:
- `mcp__plugin_linear_linear__save_issue` — create/update (description supports Markdown + Mermaid fences)
- `mcp__plugin_linear_linear__list_issues` — search by team/project
- CLI: `linear issue view SAN-NNN`, `linear issue update SAN-NNN --description-file …`
**Mermaid in MCP:** pass description as real newlines in markdown (not escaped `\n`).
---
## Views & links
| Resource | URL |
|---|---|
| CK-V2 view | https://linear.app/sanjiovani/view/v2-upgrade-30acec9f94bd |
| Team | sanjiovani |
| Conventions | [linear.md](../../../linear.md) |
---
## Don't use
- `linear-automation` / `linear-claude-skill` — archived under `.agents/skills/_archive/2026-06-11/`
- Bare issue numbers in user-facing text
- Title-only descriptions for code tasks
04 同仓库更多 Skills
mde-pr-review-supabase
Use when reviewing MDE Supabase SQL, Auth and RLS changes in a pull request.
mde-pr-review-stripe
Use when reviewing MDE Stripe payments and webhooks changes in a pull request.
mde-pr-review-nextjs
Use when reviewing MDE Next.js App Router changes in a pull request.
mde-pr-review-mastra
Use when reviewing MDE Mastra agent, tool and workflow changes in a pull request.
mde-pr-review-maps
Use when reviewing MDE Google Maps and Places changes in a pull request.
mde-pr-review-copilotkit
Use when reviewing MDE CopilotKit and AG-UI changes in a pull request.