Mermaid DiagramSAFE

🦭 A cross-device desktop AI agent with memory, autonomous goals, dynamic workflows, and headless deployment | 会记忆、能持续推进目标、会动态编排多 Agent 的跨端桌面 AI 助手,也可服务化常驻 NAS / 云端

The instruction file, verbatim from the audited commit — this is the text the model reads, and the surface the audit's instruction layer examines. Quoted here so you can judge it without cloning anything.

---
name: mermaid-diagram
description: Use when the user asks for a flowchart, architecture diagram, sequence diagram, ER diagram, state machine, gantt chart, or any visual that fits Mermaid syntax. Hope Agent's chat renders Mermaid blocks natively, so the diagram appears inline.
version: 1.0.0
author: Hope Agent
license: MIT
metadata:
  hermes:
    tags: [office, diagram, visualization, mermaid, architecture]
    related_skills: []
---

# Mermaid Diagram

## When to Use

Trigger phrases: "draw a diagram", "flowchart for ...", "sequence diagram", "architecture diagram", "ER diagram", "state machine", "gantt chart", "画一个流程图", "画个架构图".

**Mermaid is rendered natively** in Hope Agent chat (via Streamdown). Just emit a fenced ```mermaid block — the user sees the rendered SVG, no setup.

For diagrams Mermaid can't express (organic shapes, custom illustrations, hand-drawn style) consider drawing-tool skills (e.g. `excalidraw` from Hermes Agent or `drawio` from Anthropic marketplace) — those need to be installed via Quick Import.

## Pick the Right Diagram Type

| Intent | Mermaid type | Use when |
|--------|--------------|----------|
| Step-by-step process | `flowchart` | Decisions, branching, "if X then Y" |
| Time-ordered messages between actors | `sequenceDiagram` | API call traces, distributed protocols |
| State transitions of a single entity | `stateDiagram-v2` | UI state machines, workflow engines |
| Data model relationships | `erDiagram` | Database schema, entity relationships |
| Project schedule / timeline | `gantt` | Roadmaps, sprints |
| Pie / distribution | `pie` | Quick share-of-X visualization |
| Tree hierarchy (org chart, decomposition) | `flowchart TD` with subgraphs | Org / breakdown trees |
| Class hierarchy with methods | `classDiagram` | OO design, type relationships |
| User journey (steps + sentiment) | `journey` | UX flows |

If unsure, default to `flowchart`. It's the most flexible.

## Templates

### Flowchart (top-down)
```mermaid
flowchart TD
    Start([Start]) --> Decision{Condition?}
    Decision -->|Yes| ActionA[Do A]
    Decision -->|No| ActionB[Do B]
    ActionA --> End([End])
    ActionB --> End
```

### Sequence Diagram
```mermaid
sequenceDiagram
    participant U as User
    participant W as Web Server
    participant DB as Database

    U->>W: HTTP GET /api/items
    W->>DB: SELECT * FROM items
    DB-->>W: rows
    W-->>U: JSON response

    Note over U,DB: Round trip < 50ms
```

### State Diagram
```mermaid
stateDiagram-v2
    [*] --> Idle
    Idle --> Loading: fetch()
    Loading --> Success: 200 OK
    Loading --> Error: 4xx / 5xx
    Success --> Idle: reset()
    Error --> Idle: retry()
    Error --> [*]: give up
```

### ER Diagram
```mermaid
erDiagram
    USER ||--o{ ORDER : places
    ORDER ||--|{ ORDER_ITEM : contains
    PRODUCT ||--o{ ORDER_ITEM : "ordered as"
    USER {
        string id PK
        string email UK
        string name
    }
    ORDER {
        string id PK
        string user_id FK
        datetime created_at
    }
```

### Gantt
```mermaid
gantt
    title Q3 Roadmap
    dateFormat  YYYY-MM-DD
    section Auth
    Refactor middleware     :done,    a1, 2026-04-01, 2026-04-15
    OAuth migration         :active,  a2, 2026-04-15, 30d
    section Dashboard
    Search MVP              :         d1, 2026-05-01, 14d
    Analytics rewrite       :         d2, after d1, 21d
```

### Architecture (flowchart with subgraphs)
```mermaid
flowchart LR
    subgraph Frontend
        UI[React SPA]
    end
    subgraph Backend
        API[REST API]
        WS[WebSocket Hub]
    end
    subgraph Storage
        DB[(PostgreSQL)]
        Cache[(Redis)]
    end

    UI -->|HTTPS| API
    UI -->|WSS| WS
    API --> DB
    API --> Cache
    WS --> Cache
```

## Workflow

1. **Clarify intent** — if the user just says "diagram", ask via `ask_user_question`:
   - What's the diagram showing? (process / structure / relationships / timeline)
   - Who's the audience? (engineers / execs / customers)
   - Level of detail? (high-level overview / fine-grained)

2. **Pick the simplest type that fits** — flowchart > sequence > state. Don't reach for `classDiagram` if a flowchart works.

3. **Sketch nodes first, edges second** — list the entities (boxes), then connect them. Avoid dense edge spaghetti.

4. **Iterate small** — start with 5-7 nodes. Add detail only after the user confirms the structure.

5. **Verify renderability** — if you reference Mermaid syntax that's iffy (e.g. classDiagram with annotations), keep it minimal. The user will see if it doesn't render.

## Style Rules

- **One concept per diagram** — if you're tempted to add a second flow, split into two diagrams
- **Consistent shape vocabulary**:
  - `[Square]` = process / action
  - `(Round)` = entity / data
  - `{Diamond}` = decision
  - `([Stadium])` = start / end
  - `[(Cylinder)]` = database / storage
  - `((Circle))` = external system
- **Direction**: `TD` (top-down) for hierarchies, `LR` (left-right) for pipelines / data flows
- **Labels short** — node labels ≤ 4 words. Detail goes in surrounding prose.
- **Use `Note over`** in sequence diagrams to explain non-obvious behavior, not for narration

## Multi-Diagram Output

When asked for "a few diagrams" or "show me the system from different angles", output 2-3 separate ```mermaid blocks with prose between them explaining what each shows.

```
Here's the system at three levels:

**1. High-level data flow:**
\`\`\`mermaid
flowchart LR
    ...
\`\`\`

**2. Login sequence (zoom into auth):**
\`\`\`mermaid
sequenceDiagram
    ...
\`\`\`

**3. Database relationships:**
\`\`\`mermaid
erDiagram
    ...
\`\`\`
```

## Common Pitfalls

| Mistake | Fix |
|---|---|
| Single diagram with 30+ nodes | Split into 2-3 diagrams by concern |
| Mixing types (flowchart with class fields) | Pick one type; for hybrid views, do multi-diagram |
| All boxes look the same | Use shape vocabulary (decision = diamond, etc.) |
| Long node labels | Move detail to prose; keep n

SAFEgrade B · trust 89/100 Nothing in the source contradicts what it says it does. Grade A is reserved for packages that have also passed the behavioural sandbox.

What the source does

Filesystem

none-observed

Network

none-observed

Shell

none-observed

Dependencies

pinned

Secrets in source

none-found

Findings (0)

No findings outside the package's declared scope.

Gates applied: no_behavioural_pass.