dashboard-refresh

---
name: dashboard-refresh
description: Create or refresh the Claude Capabilities Dashboard. Scans agents, skills, hooks, and MCP servers from the live system and writes both the .md source and the styled HTML file. Works on first run (no existing files) or as an update.
---

# Dashboard Refresh

Create or update the Claude Code Capabilities Dashboard to reflect the current state of your setup.

## When to Use

Run `/dashboard-refresh` when:
- You want to create the dashboard for the first time (no files needed)
- You installed a new plugin, agent, or skill
- You added or removed MCP servers or hooks
- You added, removed, or rescheduled a local automation script
- You added files to `~/.claude/knowledge/`

## Target Files

| File | Role |
|---|---|
| `~/Desktop/Claude Capabilities Dashboard/claude-code-capabilities.md` | Source of truth — written first |
| `~/Desktop/Claude Capabilities Dashboard/claude-code-capabilities-styled_1.html` | Styled HTML view |

---

## Step 1 — Detect mode: Create vs Refresh

Check whether the target files exist:

```
~/.../Claude Capabilities Dashboard/claude-code-capabilities.md
~/.../Claude Capabilities Dashboard/claude-code-capabilities-styled_1.html
```

- **Both exist** → Refresh mode: read the current `.md`, diff against live system, update only what changed
- **Missing or first run** → Create mode: scan everything from scratch, write both files fresh

If the directory doesn't exist, note this — you'll need to create it with the Write tool (which creates parent dirs automatically).

---

## Step 2 — Gather live system state

Scan these sources:

### Agents
- Custom agents: read all `~/.claude/agents/*.md` — extract `name` and `description` from frontmatter
- Plugin agents: read from the system-reminder "Available agent types" list. Group by plugin prefix (`everything-claude-code:`, etc.). Strip the prefix for the display name.

### Skills
- Custom skills: read all `~/.claude/skills/*/SKILL.md` — extract `name` and `description` from frontmatter
- Built-in skills: list skills in the system-reminder that have no plugin prefix (e.g. `update-config`, `loop`, `schedule`, `simplify`, `keybindings-help`, `claude-api`)
- Plugin skills: skills in the system-reminder with a plugin prefix. Group by prefix.

### Hooks
- Read `~/.claude/settings.json` — look for the `hooks` key
- Also check `~/.claude/settings.local.json` if it exists
- For each hook, note: event type (PreToolUse, PostToolUse, PreCompact, SessionStart, Stop, SessionEnd, UserPromptSubmit), matcher/trigger, and description

### MCP Servers
- Local servers: read `~/.claude/settings.json` → `mcpServers` key. These go under "Core · Local (Developer Settings)"
- claude.ai connectors: infer from `mcp__claude_ai_<Name>__*` tool names in the system-reminder. Each unique `<Name>` is a connector. These go under "Core · claude.ai Connected Connectors"
- Plugin MCPs: look for `mcp__plugin_<plugin>_<server>__*` tool prefixes in the system-reminder

### Scripts (optional)
- Check common local automation locations: `~/.claude/cron-jobs/` (or an equivalent project-specific scripts folder your setup uses), plus any scheduler that's actually active on this machine — `launchd` on macOS (`ls ~/Library/LaunchAgents/`, `launchctl list <label>`), `crontab -l`, or `systemd --user` timers on Linux
- For each script found, note: **purpose** (what it does — read the file, don't guess), **invoked by** (which skill's instructions call it, if any — grep the skills folder for the script's filename), and **scheduled via** (the confirmed active trigger, e.g. `launchd — com.you.job-name, Mon-Fri 14:00`, or **⚠ no active trigger found** if you checked and it isn't wired up anywhere)
- **Verify scheduling claims, don't infer them from a script's own header comment.** A script's comment saying "runs daily via cron" doesn't mean it does — check the actual scheduler state (`launchctl list <label>`, not just a keyword grep over `launchctl list`; `crontab -l`; the plist file's own `StartCalendarInterval`) before reporting it as active
- If nothing is found in these locations, skip this section entirely — do not create an empty section

### Knowledge Library (optional)
- Check if `~/.claude/knowledge/` exists
- If it does: list all `.md` files excluding `README.md`
- For each file, extract: the filename (slug) and the first `# ` heading as the title
- If the folder does not exist, skip this section entirely — do not create an empty section

---

## Step 3 — Write the .md file

Write `claude-code-capabilities.md` using today's date. Structure:

```markdown
# Claude Code Capabilities
> Generated: YYYY-MM-DD

---

## Why This Dashboard Is Useful

| Reason | What it means |
|---|---|
| Single source of truth | Every agent, skill, hook, connector, and script in one place — instead of scattered across `settings.json`, plugin manifests, and memory files. |
| Onboarding / context refresh | Re-orient on what's configured without rediscovering it each session. |
| Change & drift detection | Diff this snapshot against the last one to see what was added, removed, or changed. |
| First-pass breach/incident checklist | A starting point for what has access and what's auto-approved — pair with your own credentials/permissions audit for a full security review. |
| Debugging aid | Confirm a hook, skill, or trigger is actually registered before assuming the bug lives elsewhere. |
| Cleanup candidate list | Surface stale or unused agents and skills worth pruning. |
| Shareable snapshot | Hand to a collaborator or incident responder without exposing actual secret values. |
| Automation audit trail | Every cron job, launchd agent, and RemoteTrigger routine that runs unattended — with confirmed, not assumed, scheduling status. |

---

## Agents

### Core (Custom)
| Agent | Description |
|---|---|
| `name` | description |

### Plugin: everything-claude-code (v1.9.0)
| Agent | Description |
|---|---|
...

> Invoke agents with `/agent:<name>` or via the Agent tool in conversation.

---

## Skills

### Core (Built-in)
| Skill | Description |
|---|---|
...

### Core (Custom)
| Skill | Description |
|---|---|
...

### Plugin: claude-mem (v10.6.3)
| Skill | Description |
|---|---|
...

### Plugin: everything-claude-code (v1.9.0)
#### Core Workflow
| Skill | Description |
...

> Invoke with `/skill-name` or `/plugin:skill-name`

---

## Hooks

### Plugin: everything-claude-code (v1.9.0)

#### PreToolUse
| Hook | Trigger | Description |
|---|---|---|
...

#### PostToolUse
...

#### Stop
...

#### SessionStart / SessionEnd / PreCompact
...

---

### Plugin: claude-mem (v10.6.3)
...

---

## MCP Servers

### Core (Local — Developer Settings)
| Name | Description |
|---|---|
...

### Core (claude.ai Connected Connectors)
| Name | Description |
|---|---|
...

### Plugin: everything-claude-code (v1.9.0)
| MCP Server | Key Tools |
|---|---|
...

---

## Scripts
> Only include this section if you found scripts per Step 2. Group by location.

### `~/.claude/cron-jobs/` — scheduled maintenance scripts
| Script | Purpose | Scheduled via |
|---|---|---|
| `script-name.sh` | What it does | `launchd` — `com.you.label`, schedule (confirmed active) — or **⚠ no active trigger found** |

> Add one table per script location you scanned, and a separate table for any relevant purpose/invoked-by columns your skills scripts use (e.g. `| Script | Purpose | Invoked by |` for scripts a skill calls rather than a scheduler).

## Knowledge Library
> Only include this section if `~/.claude/knowledge/` exists and contains .md files.

| File | Title |
|---|---|
| `filename` | First # heading from the file |

> Location: `~/.claude/knowledge/`

---

## Scope & Limitations

⚠️ **Single-machine, point-in-time snapshot** (as of the date above). The Scripts section lists *registered* RemoteTrigger/Cowork cloud routines if you have any, but not ad-hoc Cowork tasks or what happens inside a Cowork sandbox — and it doesn't cover other machines or anything added since the last `/dashboard-refresh`. This dashboard also does not audit credentials, secrets, or permission auto-approvals — it's a capability inventory, not a security audit. Useful as a **first checklist**, not a complete source of truth on its own.
```

---

## Step 4 — Write the HTML file

Write the full HTML file. In **Refresh mode**, preserve the existing CSS/JS and only replace the `<div class="wrapper">...</div>` block. In **Create mode**, write the complete file from the template below.

### HTML Template

The complete file structure to use (or preserve) for the HTML:

```html
<!DOCTYPE html>
<html lang="en">
<head>
<meta charset="UTF-8">
<meta name="viewport" content="width=device-width, initial-scale=1.0">
<title>Claude Code Capabilities</title>
<link href="https://fonts.googleapis.com/css2?family=DM+Mono:wght@300;400;500&family=Syne:wght@400;600;700;800&family=DM+Sans:wght@300;400;500&display=swap" rel="stylesheet">
<style>
  :root {
    --bg: #0a0a0f; --surface: #111118; --surface2: #1a1a24; --border: #2a2a3a;
    --accent: #7c6bff; --accent2: #ff6b9d; --accent3: #6bffd8; --accent4: #ffd96b;
    --text: #e8e8f0; --muted: #6b6b80;
    --tag-bg: rgba(124,107,255,0.12); --tag-border: rgba(124,107,255,0.3);
    --nav-bg: rgba(10,10,15,0.88);
    --h1-gradient: linear-gradient(135deg,#fff 0%,rgba(255,255,255,0.6) 100%);
    --table-cell: #c8c8d8;
  }
  [data-theme="light"] {
    --bg: #f4f4f9; --surface: #ffffff; --surface2: #ebebf4; --border: #d8d8e8;
    --accent: #6b5ce7; --accent2: #e8547a; --accent3: #00a87a; --accent4: #c98a00;
    --text: #1a1a2e; --muted: #7878a0;
    --tag-bg: rgba(107,92,231,0.1); --tag-border: rgba(107,92,231,0.25);
    --nav-bg: rgba(244,244,249,0.92);
    --h1-gradient: linear-gradient(135deg,#1a1a2e 0%,#4a4a7a 100%);
    --table-cell: #4a4a6a;
  }
  @media (prefers-color-scheme: light) {
    :root:not([data-theme="dark"]):not([data-theme="light"]) {
      --bg: #f4f4f9; --surface: #ffffff; --surface2: #ebebf4; --border: #d8d8e8;
      --accent: #6b5ce7; --accent2: #e8547a; --accent3: #00a87a; --accent4: #c98a00;
      --text: #1a1a2e; --muted: #7878a0;
      --tag-bg: rgba(107,92,231,0.1); --tag-border: rgba(107,92,231,0.25);
      --nav-bg: rgba(244,244,249,0.92);
      --h1-gradient: linear-gradient(135deg,#1a1a2e 0%,#4a4a7a 100%);
      --table-cell: #4a4a6a;
    }
  }
  * { margin:0; padding:0; box-sizing:border-box; }
  html { scroll-behavior: smooth; }
  body { background:var(--bg); color:var(--text); font-family:'DM Sans',sans-serif; font-weight:300; line-height:1.6; min-height:100vh; overflow-x:hidden; }
  body::before { content:''; position:fixed; top:-50%; left:-50%; width:200%; height:200%;
    background: radial-gradient(ellipse 60% 40% at 20% 20%,rgba(124,107,255,0.08) 0%,transparent 60%),
                radial-gradient(ellipse 50% 35% at 80% 70%,rgba(107,255,216,0.06) 0%,transparent 60%),
                radial-gradient(ellipse 40% 30% at 60% 10%,rgba(255,107,157,0.05) 0%,transparent 60%);
    pointer-events:none; z-index:0; }
  .wrapper { position:relative; z-index:1; max-width:1100px; margin:0 auto; padding:0 32px 80px; }
  header { padding:72px 0 56px; border-bottom:1px solid var(--border); margin-bottom:64px;
    display:flex; align-items:flex-end; justify-content:space-between; gap:32px; flex-wrap:wrap; }
  .eyebrow { font-family:'DM Mono',monospace; font-size:11px; letter-spacing:0.18em; text-transform:uppercase;
    color:var(--accent); margin-bottom:16px; display:flex; align-items:center; gap:10px; }
  .eyebrow::before { content:''; display:inline-block; width:24px; height:1px; background:var(--accent); }
  h1 { font-family:'Syne',sans-serif; font-size:clamp(36px,6vw,64px); font-weight:800;
    letter-spacing:-0.02em; line-height:1.05; background:var(--h1-gradient);
    -webkit-background-clip:text; -webkit-text-fill-color:transparent; background-clip:text; }
  .header-meta { font-family:'DM Mono',monospace; font-size:12px; color:var(--muted); text-align:right; line-height:2; }
  .header-meta strong { color:var(--text); font-weight:500; }
  nav { position:sticky; top:0; z-index:100; background:var(--nav-bg); backdrop-filter:blur(20px);
    border-bottom:1px solid var(--border); margin:0 -32px 64px; padding:0 32px; }
  .nav-inner { display:flex; gap:0; overflow-x:auto; scrollbar-width:none; }
  .nav-inner::-webkit-scrollbar { display:none; }
  nav a { font-family:'DM Mono',monospace; font-size:11px; letter-spacing:0.12em; text-transform:uppercase;
    color:var(--muted); text-decoration:none; padding:16px 20px; white-space:nowrap;
    border-bottom:2px solid transparent; transition:color 0.2s, border-color 0.2s; }
  nav a:hover { color:var(--text); border-bottom-color:var(--accent); }
  section { margin-bottom:80px; }
  .section-header { display:flex; align-items:center; gap:16px; margin-bottom:40px;
    padding-bottom:20px; border-bottom:1px solid var(--border); }
  .section-icon { width:40px; height:40px; border-radius:10px; display:flex; align-items:center;
    justify-content:center; font-size:18px; flex-shrink:0; }
  .icon-agents     { background:linear-gradient(135deg,rgba(124,107,255,0.3),rgba(124,107,255,0.1)); }
  .icon-skills     { background:linear-gradient(135deg,rgba(107,255,216,0.3),rgba(107,255,216,0.1)); }
  .icon-hooks      { background:linear-gradient(135deg,rgba(255,217,107,0.3),rgba(255,217,107,0.1)); }
  .icon-mcp        { background:linear-gradient(135deg,rgba(255,107,157,0.3),rgba(255,107,157,0.1)); }
  .icon-scripts    { background:linear-gradient(135deg,rgba(124,107,255,0.25),rgba(255,217,107,0.15)); }
  .icon-knowledge  { background:linear-gradient(135deg,rgba(107,255,216,0.2),rgba(124,107,255,0.2)); }
  h2 { font-family:'Syne',sans-serif; font-size:28px; font-weight:700; letter-spacing:-0.01em; }
  .section-count { margin-left:auto; font-family:'DM Mono',monospace; font-size:11px; color:var(--muted); letter-spacing:0.08em; }
  .subsection { margin-bottom:40px; }
  .subsection-label { font-family:'DM Mono',monospace; font-size:10px; letter-spacing:0.2em; text-transform:uppercase;
    color:var(--muted); margin-bottom:12px; display:flex; align-items:center; gap:8px; }
  .subsection-label::after { content:''; flex:1; height:1px; background:var(--border); }
  h3 { font-family:'Syne',sans-serif; font-size:15px; font-weight:600; color:var(--muted); margin-bottom:16px; }
  h4 { font-family:'Syne',sans-serif; font-size:13px; font-weight:600; color:var(--accent);
    margin:24px 0 10px; text-transform:uppercase; letter-spacing:0.08em; }
  .card-grid { display:grid; grid-template-columns:repeat(auto-fill,minmax(280px,1fr)); gap:10px; margin-bottom:16px; }
  .card { background:var(--surface); border:1px solid var(--border); border-radius:10px;
    padding:14px 16px; display:flex; flex-direction:column; gap:6px;
    transition:border-color 0.2s, background 0.2s; cursor:default; }
  .card:hover { border-color:rgba(124,107,255,0.4); background:var(--surface2); }
  .card-name { font-family:'DM Mono',monospace; font-size:12px; font-weight:500; color:var(--accent);
    white-space:nowrap; overflow:hidden; text-overflow:ellipsis; }
  /* color convention: agents = default purple, skills = green, knowledge = gold */
  /* agent `color` frontmatter does NOT map to card color — all agents use plain card-name (purple) */
  .card-name.green { color:var(--accent3); }
  .card-name.pink  { color:var(--accent2); }
  .card-name.gold  { color:var(--accent4); }
  .card-desc { font-size:12px; color:var(--muted); line-height:1.5; }
  .hook-table { width:100%; border-collapse:collapse; font-size:12.5px; margin-bottom:8px; }
  .hook-table thead tr { border-bottom:1px solid var(--border); }
  .hook-table th { font-family:'DM Mono',monospace; font-size:10px; letter-spacing:0.14em;
    text-transform:uppercase; color:var(--muted); text-align:left; padding:8px 12px 10px; font-weight:400; }
  .hook-table td { padding:10px 12px; border-bottom:1px solid rgba(42,42,58,0.5);
    vertical-align:top; color:var(--table-cell); }
  .hook-table tr:last-child td { border-bottom:none; }
  .hook-table tr:hover td { background:var(--surface2); }
  .hook-table td:first-child { font-weight:500; color:var(--text); white-space:nowrap; }
  .trigger-badge { display:inline-block; font-family:'DM Mono',monospace; font-size:10px;
    background:var(--tag-bg); border:1px solid var(--tag-border); color:var(--accent);
    border-radius:4px; padding:2px 6px; white-space:nowrap; }
  .trigger-badge.gold { background:rgba(255,217,107,0.1); border-color:rgba(255,217,107,0.3); color:var(--accent4); }
  .mcp-grid { display:grid; grid-template-columns:repeat(auto-fill,minmax(320px,1fr)); gap:14px; }
  .mcp-card { background:var(--surface); border:1px solid var(--border); border-radius:12px;
    padding:20px; transition:border-color 0.2s; }
  .mcp-card:hover { border-color:rgba(255,107,157,0.4); }
  .mcp-name { font-family:'Syne',sans-serif; font-size:15px; font-weight:700; margin-bottom:10px; color:var(--text); }
  .mcp-tools { display:flex; flex-wrap:wrap; gap:5px; }
  .tool-chip { font-family:'DM Mono',monospace; font-size:10px; background:rgba(255,107,157,0.1);
    border:1px solid rgba(255,107,157,0.25); color:var(--accent2); border-radius:4px; padding:3px 7px; }
  .invoke-hint { margin-top:24px; padding:14px 18px; background:var(--surface);
    border:1px solid var(--border); border-radius:8px; font-family:'DM Mono',monospace;
    font-size:12px; color:var(--muted); }
  .invoke-hint code { color:var(--accent3); background:rgba(107,255,216,0.1); border-radius:4px; padding:2px 6px; }
  .phase-label { font-family:'DM Mono',monospace; font-size:10px; letter-spacing:0.16em;
    text-transform:uppercase; display:inline-flex; align-items:center; gap:6px;
    margin:28px 0 12px; color:var(--accent4); }
  .phase-label::before { content:''; width:6px; height:6px; border-radius:50%; background:var(--accent4); }
  .theme-toggle { display:flex; align-items:center; gap:2px; background:var(--surface);
    border:1px solid var(--border); border-radius:8px; padding:3px; margin-top:12px; }
  .theme-btn { font-family:'DM Mono',monospace; font-size:10px; letter-spacing:0.08em; color:var(--muted);
    background:none; border:none; border-radius:5px; padding:5px 10px; cursor:pointer;
    transition:background 0.15s, color 0.15s; white-space:nowrap; }
  .theme-btn:hover { color:var(--text); }
  .theme-btn.active { background:var(--surface2); color:var(--accent); }
  @keyframes fadeUp { from { opacity:0; transform:translateY(20px); } to { opacity:1; transform:translateY(0); } }
  section { animation:fadeUp 0.5s ease both; }
  section:nth-child(1) { animation-delay:0.05s; }
  section:nth-child(2) { animation-delay:0.12s; }
  section:nth-child(3) { animation-delay:0.19s; }
  section:nth-child(4) { animation-delay:0.26s; }
</style>
</head>
<body>
<div class="wrapper">

  <header>
    <div class="header-left">
      <div class="eyebrow">Configuration Export</div>
      <h1>Claude Code<br>Capabilities</h1>
    </div>
    <div class="header-meta">
      <strong>Generated</strong> YYYY-MM-DD<br>
      <strong>Agents</strong> N · <strong>Skills</strong> N<br>
      <strong>Hooks</strong> N · <strong>MCP Servers</strong> N · <strong>Scripts</strong> N
      <div class="theme-toggle">
        <button class="theme-btn" data-mode="dark" onclick="setTheme('dark')">🌙 Dark</button>
        <button class="theme-btn" data-mode="light" onclick="setTheme('light')">☀️ Light</button>
        <button class="theme-btn" data-mode="system" onclick="setTheme('system')">💻 System</button>
      </div>
    </div>
  </header>

  <nav>
    <div class="nav-inner">
      <a href="#why">Why This</a>
      <a href="#agents">Agents</a>
      <a href="#skills">Skills</a>
      <a href="#hooks">Hooks</a>
      <a href="#mcp">MCP Servers</a>
      <a href="#scripts">Scripts</a>
      <a href="#knowledge">Knowledge</a>
      <a href="#scope">Scope</a>
    </div>
  </nav>

  <!-- WHY THIS DASHBOARD SECTION — always render, right after </nav>, before Agents -->
  <section id="why">
    <div class="section-header">
      <div class="section-icon icon-scripts">💡</div>
      <h2>Why This Dashboard Is Useful</h2>
    </div>
    <div class="card-grid">
      <div class="card"><div class="card-name">Single source of truth</div><div class="card-desc">Every agent, skill, hook, connector, and script in one place — instead of scattered across settings.json, plugin manifests, and memory files.</div></div>
      <div class="card"><div class="card-name">Onboarding / context refresh</div><div class="card-desc">Re-orient on what's configured without rediscovering it each session.</div></div>
      <div class="card"><div class="card-name">Change &amp; drift detection</div><div class="card-desc">Diff this snapshot against the last one to see what was added, removed, or changed.</div></div>
      <div class="card"><div class="card-name">First-pass breach/incident checklist</div><div class="card-desc">A starting point for what has access and what's auto-approved — pair with your own credentials/permissions audit for a full security review.</div></div>
      <div class="card"><div class="card-name">Debugging aid</div><div class="card-desc">Confirm a hook, skill, or trigger is actually registered before assuming the bug lives elsewhere.</div></div>
      <div class="card"><div class="card-name">Cleanup candidate list</div><div class="card-desc">Surface stale or unused agents and skills worth pruning.</div></div>
      <div class="card"><div class="card-name">Shareable snapshot</div><div class="card-desc">Hand to a collaborator or incident responder without exposing actual secret values.</div></div>
      <div class="card"><div class="card-name">Automation audit trail</div><div class="card-desc">Every cron job, launchd agent, and RemoteTrigger routine that runs unattended — with confirmed, not assumed, scheduling status.</div></div>
    </div>
  </section>

  <!-- AGENTS SECTION -->
  <section id="agents">
    <div class="section-header">
      <div class="section-icon icon-agents">🤖</div>
      <h2>Agents</h2>
      <span class="section-count">N agents</span>
    </div>

    <!-- one .subsection per group, e.g.: -->
    <div class="subsection">
      <div class="subsection-label">Core · Custom</div>
      <div class="card-grid">
        <div class="card">
          <div class="card-name">agent-name</div>
          <div class="card-desc">description</div>
        </div>
      </div>
    </div>

    <div class="subsection">
      <div class="subsection-label">Plugin · everything-claude-code v1.9.0</div>
      <div class="card-grid">
        <!-- one .card per agent -->
      </div>
    </div>

    <div class="invoke-hint">Invoke agents with <code>/agent:&lt;name&gt;</code> or via the Agent tool in conversation.</div>
  </section>

  <!-- SKILLS SECTION -->
  <section id="skills">
    <div class="section-header">
      <div class="section-icon icon-skills">⚡</div>
      <h2>Skills</h2>
      <span class="section-count">N skills</span>
    </div>

    <div class="subsection">
      <div class="subsection-label">Core · Built-in</div>
      <div class="card-grid">
        <!-- card-name green for skills -->
        <div class="card"><div class="card-name green">skill-name</div><div class="card-desc">description</div></div>
      </div>
    </div>

    <div class="subsection">
      <div class="subsection-label">Core · Custom</div>
      <div class="card-grid">
        <!-- each custom skill gets its own .card — never nest subsection-label inside card-grid -->
        <div class="card"><div class="card-name green">skill-name</div><div class="card-desc">description</div></div>
      </div>
    </div>

    <!-- plugin skills: use h4 for sub-categories within a plugin -->
    <div class="subsection">
      <div class="subsection-label">Plugin · everything-claude-code v1.9.0</div>
      <h4>Core Workflow</h4>
      <div class="card-grid"><!-- cards --></div>
      <h4>Code Quality &amp; Review</h4>
      <div class="card-grid"><!-- cards --></div>
      <!-- etc. -->
    </div>

    <div class="invoke-hint">Invoke with <code>/skill-name</code> or <code>/plugin:skill-name</code></div>
  </section>

  <!-- HOOKS SECTION -->
  <section id="hooks">
    <div class="section-header">
      <div class="section-icon icon-hooks">🪝</div>
      <h2>Hooks</h2>
      <span class="section-count">N hooks</span>
    </div>

    <div class="subsection">
      <div class="subsection-label">Plugin · everything-claude-code v1.9.0</div>

      <div class="phase-label">PreToolUse</div>
      <table class="hook-table">
        <thead><tr><th>Hook</th><th>Trigger</th><th>Description</th></tr></thead>
        <tbody>
          <tr><td>Hook name</td><td><span class="trigger-badge">Bash</span></td><td>description</td></tr>
        </tbody>
      </table>

      <div class="phase-label">PostToolUse</div>
      <table class="hook-table"><!-- ... --></table>

      <div class="phase-label">Stop</div>
      <table class="hook-table"><!-- ... --></table>

      <!-- lifecycle hooks use .gold badge -->
      <div class="phase-label">SessionStart / SessionEnd / PreCompact</div>
      <table class="hook-table">
        <thead><tr><th>Hook</th><th>Phase</th><th>Description</th></tr></thead>
        <tbody>
          <tr><td>Hook name</td><td><span class="trigger-badge gold">SessionStart</span></td><td>description</td></tr>
        </tbody>
      </table>
    </div>
  </section>

  <!-- MCP SECTION -->
  <section id="mcp">
    <div class="section-header">
      <div class="section-icon icon-mcp">🔌</div>
      <h2>MCP Servers</h2>
      <span class="section-count">N servers</span>
    </div>

    <div class="subsection">
      <div class="subsection-label">Core · Local (Developer Settings)</div>
      <div class="mcp-grid">
        <div class="mcp-card">
          <div class="mcp-name">Server Name</div>
          <div class="card-desc" style="margin-bottom:12px">description</div>
          <div class="mcp-tools">
            <span class="tool-chip">tool-name</span>
          </div>
        </div>
      </div>
    </div>

    <div class="subsection">
      <div class="subsection-label">Core · claude.ai Connected Connectors</div>
      <div class="mcp-grid"><!-- mcp-cards --></div>
    </div>

    <div class="subsection">
      <div class="subsection-label">Plugin · everything-claude-code v1.9.0</div>
      <div class="mcp-grid"><!-- mcp-cards --></div>
    </div>
  </section>

  <!-- SCRIPTS SECTION — only render if scripts were found per Step 2 -->
  <section id="scripts">
    <div class="section-header">
      <div class="section-icon icon-scripts">📜</div>
      <h2>Scripts</h2>
      <span class="section-count">N scripts</span>
    </div>

    <p style="font-size:13px;color:var(--muted);margin-bottom:24px;max-width:720px">Scripts that run outside the conversational tool-call model — invoked by a skill's instructions, or scheduled to run unattended.</p>

    <div class="subsection">
      <div class="subsection-label">~/.claude/cron-jobs/</div>
      <table class="hook-table">
        <thead><tr><th>Script</th><th>Purpose</th><th>Scheduled via</th></tr></thead>
        <tbody>
          <tr><td>script-name.sh</td><td>What it does</td><td>launchd — com.you.label, schedule (confirmed active)</td></tr>
        </tbody>
      </table>
    </div>
  </section>

  <!-- KNOWLEDGE SECTION — only render if ~/.claude/knowledge/ exists and has .md files -->
  <section id="knowledge">
    <div class="section-header">
      <div class="section-icon icon-knowledge">📚</div>
      <h2>Knowledge Library</h2>
      <span class="section-count">N notes</span>
    </div>

    <div class="subsection">
      <div class="subsection-label">~/.claude/knowledge</div>
      <div class="card-grid">
        <div class="card">
          <div class="card-name gold">filename-slug</div>
          <div class="card-desc">Title extracted from first # heading</div>
        </div>
      </div>
    </div>

    <div class="invoke-hint">Explanatory notes on how Claude Code works — concepts, comparisons, mental models. Not rules (memory) or prompts (skills).</div>
  </section>

  <!-- SCOPE & LIMITATIONS — always render, at the very end, after Knowledge -->
  <section id="scope">
    <div class="section-header">
      <div class="section-icon icon-scripts">⚠️</div>
      <h2>Scope &amp; Limitations</h2>
    </div>
    <p style="font-size:13px;color:var(--table-cell);line-height:1.6;max-width:720px">
      <strong style="color:var(--text)">Single-machine, point-in-time snapshot</strong> (as of the date above). The Scripts section lists <em>registered</em> RemoteTrigger/Cowork cloud routines if you have any, but not ad-hoc Cowork tasks or what happens inside a Cowork sandbox — and it doesn't cover other machines or anything added since the last <code style="background:rgba(255,107,157,0.12); border-radius:4px; padding:1px 5px;">/dashboard-refresh</code>. This dashboard also does not audit credentials, secrets, or permission auto-approvals — it's a capability inventory, not a security audit. Useful as a <strong style="color:var(--text)">first checklist</strong>, not a complete source of truth on its own.
    </p>
  </section>

</div>

<script>
  function setTheme(mode) {
    if (mode === 'system') {
      document.documentElement.removeAttribute('data-theme');
      localStorage.removeItem('theme');
    } else {
      document.documentElement.setAttribute('data-theme', mode);
      localStorage.setItem('theme', mode);
    }
    document.querySelectorAll('.theme-btn').forEach(btn => {
      btn.classList.toggle('active', btn.dataset.mode === mode);
    });
  }
  (function() {
    const saved = localStorage.getItem('theme');
    const mode = saved || 'system';
    if (saved) document.documentElement.setAttribute('data-theme', saved);
    document.querySelectorAll('.theme-btn').forEach(btn => {
      btn.classList.toggle('active', btn.dataset.mode === mode);
    });
  })();
</script>
</body>
</html>
```

Fill in the placeholder content (`N`, `YYYY-MM-DD`, card lists, etc.) from the data gathered in Step 2.

---

## Step 5 — Counts & summary line

Update the header stats line in the HTML:
```html
<strong>Agents</strong> 26 · <strong>Skills</strong> 130+<br>
<strong>Hooks</strong> 23 · <strong>MCP Servers</strong> 9 · <strong>Scripts</strong> 5
```
Count actual items from Step 2. Use `N+` notation if a section has many items from a plugin (e.g. `130+` skills). Omit `· Scripts N` from the header entirely if the Scripts section wasn't included (nothing found in Step 2).

---

## Step 6 — Verify

After writing both files confirm:
- Date updated to today in both files, including the Scope & Limitations section (which restates "as of the date above")
- Counts in HTML header match actual totals
- All agents/skills/hooks/MCPs from Step 2 are present
- "Why This Dashboard Is Useful" section present near the top (right after nav) and Scope & Limitations section present at the very bottom — these are fixed content, always render both regardless of mode
- Knowledge section present if `~/.claude/knowledge/` exists, absent if it doesn't
- Scripts section present only if you actually found scripts in Step 2, and every "scheduled via" claim is backed by a real check, not a script comment
- In Refresh mode: no items removed unless they genuinely no longer exist

---

## Notes

- Plugin versions can be found in the plugin's package.json or manifest. If unknown, omit the version.
- "Core · Local (Developer Settings)" entries must be maintained manually — do not auto-remove them.
- claude.ai connectors come from `mcp__claude_ai_<Name>__*` tool prefixes. Each unique `<Name>` is one connector.
- If the target folder doesn't exist, the Write tool will create it automatically.
- Knowledge Library section is optional and plugin-independent — include it only if `~/.claude/knowledge/` exists. Any user can create this folder to get the section.
- Scripts section is optional — include it only if Step 2 actually found scripts. Don't trust a script's own header comment about how/when it runs; confirm against the live scheduler state.
- "Why This Dashboard Is Useful" (top, right after nav) and "Scope & Limitations" (bottom, after Knowledge) are both fixed, always-rendered content — not derived from Step 2 scanning, and not optional. Be precise about what's actually covered: if you added a Scripts section listing RemoteTrigger/Cowork routines, don't let the disclaimer claim Cowork is excluded entirely — scope it to what's genuinely missing (ad-hoc Cowork tasks, sandbox internals, other machines, changes since last refresh). This template intentionally omits credentials/secrets and permissions auditing — say so plainly rather than implying broader security coverage than the dashboard actually provides.