0001 How do I add a new project?

1. Look at the left navigation sidebar and click on the 📁 Workspaces tab.
2. In the top right corner of the Workspaces view, locate and click the bright ＋ New Workspace button.
3. A modal will pop up prompting you for a Workspace Name (e.g., "Tung Shing App") and an optional Description.
4. Fill out the details and click the Create Workspace button to initialize your new project environment.

0002 How do I toggle the Auto-Reply Engine?

1. Navigate to the 🛰️ Mission Control view from your left sidebar.
2. Scroll down past the Telemetry Overview to the Swarm Protocols & Overrides section.
3. Locate the cyan-colored Auto-Reply Engine card.
4. Click anywhere on the card to toggle the engine. A toast notification will appear in the bottom right confirming if it was Enabled or Disabled.

0003 How do I switch Swarm Profiles?

1. Look at the top left of your sidebar, just below the Clawpy logo.
2. You will see an "ACTIVE PROFILE" dropdown menu (usually set to 'default').
3. Click the dropdown menu and select the profile you wish to switch to.
4. The dashboard will automatically sync and reboot the workspace into the selected Profile context.

0004 How do I execute a GWS command?

1. Ensure you are in the 🛰️ Mission Control view.
2. Scroll down to the Swarm Protocols & Overrides panel.
3. Locate the green GWS Console card and click it.
4. A prompt window will appear. Enter your Global Workspace System command (for example: gws gmail list).
5. Press OK. The system will execute the command via the Workspace Adapter and return the JSON result directly to you.

0005 How do I view telemetry?

1. Select the 🛰️ Mission Control view from the sidebar.
2. You will see active Swarm Models, total costs, and token consumption directly in the Telemetry section.
3. To customize exactly which telemetry panels show up, click the ⚙️ Configure button in the top right.
4. Toggle options like "Show Swarm Cost" and "Show Semantic Memory Size" to curate your dashboard.

0006 How do I query Semantic Memory?

1. Select the 🧠 Memory view from the sidebar.
2. Locate the "Semantic Memory Check" search bar at the top of the interface.
3. Type in a query to search through the vector store for historical context, tool results, and schemas.
4. Click 🔍 Search array the Layer 3 Vector database directly from your current workspace context.

0007 How do I add a Secret to my Vault?

1. Navigate to the 🔐 Vault section.
2. At the top of the Vault hero section, look for the 'Add New Secret' input fields.
3. Enter the Key name in the first box (e.g., MINIMAX_API_KEY).
4. Enter the Secret value in the second obscured box.
5. Click the ＋ Add button to securely encrypt it into the vault assigned to your active profile.

0008 How do I import a Community Skill?

1. Select the ⚡️ Skills view from the sidebar.
2. At the top of the Skills grid, look for the ↓ Import bar.
3. Paste the URL of a GitHub directory containing a Clawpy Community Skill.
4. Press the ↓ Import button. The backend will scrape the files securely and unzip them directly into your workflow engine.

0009 How do I review a Swarm Debate?

1. Click into the 💬 Debate UI tab.
2. Type a complex technical topic or dilemma into the "Debate Topic" textbox.
3. Check the boxes for which specialized agents you want involved (MainAgent, Architect, Editor).
4. Click ▶ Start Debate. The background swarm will argue the topic and output a final consensus log.

0010 How do I edit my Stack Config?

1. Navigate to the 🔐 Archetype and Vault view.
2. You will see a list of logical LLM nodes representing different roles (Architect, Builder, Implementor, Auditor).
3. Update the string identifiers with your preferred models (e.g., gpt-5 or claude-4-sonnet).
4. Click Save Stack Profile to permanently commit this model array mapping to your Vault.

0011 How do I export or migrate data?

1. On almost any dashboard page, look to the very top right navigation utility bar.
2. Click 💾 Export to zip all of your configurations, vaults, and memory states into a transferable payload.
3. Alternatively, click 🚀 Migrate OpenClaw to select a local directory to port legacy '.env' files, skills, and memory logs into your modern Clawpy project.

0012 How do I review the Agent Soul?

1. Click into the ✨ Soul tab on the sidebar navigation.
2. This page accesses the soul.md system prompt component which guides the core persona.
3. Review the raw Markdown dump to ensure your agent's underlying behavioral logic matches its directives.

0013 How do I check Heartbeats?

1. Navigate to the ❤️ Heartbeat view.
2. The real-time server clock will pulse, accompanied by the contents of heartbeat.md.
3. Review the logged health pings to analyze uptime and internal autonomous checkpoints across your docker arrays.

0014 How do I browse Starter Kits?

1. Navigate to the 📦 Kits view located in the sidebar options.
2. This grid will surface any pre-configured template bundles located in the /kits system directory.
3. Each block contains relevant metadata ensuring fast scaffolding for specific industry workflows.

0015 How do I browse Starter Kits?

1. Open the 📦 Starter Kits menu from the sidebar.
2. Browse the full catalog of pre-built agent swarm configurations — from SaaS builders to marketing empires.
3. Click Activate on any kit to trigger the setup wizard, which will guide you through required API keys and integrations.

0016 How do I use Spec-Driven Development (GitHub Spec Kit)?

1. Create a .speckit/ folder in your active workspace directory.
2. Write a constitution.md file containing your high-level architectural rules and project requirements.
3. Clawpy agents will automatically detect this configuration on boot and inject your rules into their system prompts for strict enforcement.
4. For more advanced details, refer to the full SDD documentation →

0017 How do I connect an MCP tool server?

1. Navigate to the ⚡️ Skills view located in the sidebar options.
2. At the top of the interface next to the ↓ Import bar, click the 🔌 Connect MCP button.
3. A modal will appear prompting you to define the connection command (e.g., npx or uvx) and its arguments. For instance, command npx and arguments -y @modelcontextprotocol/server-postgres postgresql://localhost/mydb.
4. Click Connect to store the configuration. Clawpy will spawn the server natively and proxy its tools over standard input/output directly to your agents!

0018 How do I configure CEO Autonomy? (Clawpy vs OpenCLAW)

1. Navigate to the 🐝 Agent Swarm Manager view.
2. Locate the card for the "ceo" agent.
3. Click ⚙️ Edit Config on the ceo card.
4. Use the Autonomy Level (1-10) slider to set boundaries.
   • Level 10: Full Autonomy - Acts independently on all tasks.
   • Level 5: Sync - Requests human approval at feature milestones.
   • Level 1: Zero Autonomy - Requests permission before executing any changes.
5. By abstracting this control to the CEO, the rest of the swarm remains fast and autonomous, blocked only when the CEO hits its strategic/spending boundaries and must ask for your permission.

0019 How do I set limits on sub-agents?

1. Navigate to the 🐝 Agent Swarm page in the dashboard sidebar.
2. Locate the ⚙️ Swarm Autonomy Limits panel displayed above the division panels. This is where you control worker spawning limits and budgets.
3. Configure the limits:
   • Max Workers per Division — Simultaneous workers per division head. Default: 3
   • Max Workers (System-wide) — Total concurrent workers. Default: 6
   • Worker Budget (Hourly) — Per-worker hourly cap. Default: $2.00
   • Worker Budget (Daily) — Combined daily budget. Default: $10.00
   • Worker TTL — Auto-termination timer. Default: 30 min
   • Worker Model — LLM for workers. Default: gpt-4o-mini
4. Set the Approval Mode: Always (every spawn needs approval), Over Budget ★ (recommended), or Never (fully autonomous).
5. Check or uncheck the Allowed Spawners boxes (CEO, CRO, CTO, CMO).
6. Click 💾 Save Configuration to apply your limits.

0020 How do I create a Workspace, Swarm Space, Division(s), and Agent(s)?

Build your organization top-down: create a workspace, add swarmspaces, prepare reusable agent cards, then assemble those agents into divisions and reporting hierarchies.

1. Create a Workspace in the left sidebar Workspaces page (+ New Workspace).
2. Open Agent Swarm → Open Swarm Configurator, then in Swarmspaces create one or more swarm spaces.
3. In Agents, create your agent cards/templates.
4. In Divisions, select a swarmspace, create division(s), set a division head, then add/remove agents into each division.
5. In Hierarchy, set reporting links between division agents.

1) Workspaces (max 30)
      |
      v
2) Swarmspaces (max 30 per Workspace)
      |
      v
3) Agents (create templates/cards)
      |
      v
4) Divisions (max 30 per Swarmspace)
  + Agent assignments (max 30 Agents per Division)
      |
      v
5) Hierarchy links (reports-to between division agents)