fatby/dd-os-data

DunCrew Data Provider

Provides structured data from OpenClaw's configuration files for DunCrew frontend visualization.

What This Skill Does

When DunCrew frontend requests data, this skill instructs the agent to:

• Read SOUL.md — Extract soul identity, core truths, boundaries, and vibe
• Read IDENTITY.md — Extract agent identity information
• List Skills — Scan skills directory and return all installed skills
• Read Memories — Extract memory entries from memory directory

Usage

When the user or DunCrew requests data, respond with the appropriate structured format.

Get Soul Data

When asked for soul data, read SOUL.md and IDENTITY.md, then return:

{
  "type": "ddos.soul",
  "data": {
    "identity": {
      "name": "<from IDENTITY.md or SOUL.md>",
      "essence": "<subtitle from SOUL.md>",
      "symbol": "<emoji if any>"
    },
    "coreTruths": [
      {
        "id": "truth-0",
        "title": "<short title>",
        "principle": "<the principle statement>",
        "description": "<explanation>"
      }
    ],
    "boundaries": [
      {
        "id": "boundary-0",
        "rule": "<boundary rule text>"
      }
    ],
    "vibeStatement": "<vibe section content>",
    "continuityNote": "<continuity section content>",
    "rawContent": "<full SOUL.md content>"
  }
}

Get Skills List

When asked for skills, scan the skills directories and return:

{
  "type": "ddos.skills",
  "data": {
    "skills": [
      {
        "name": "<skill folder name>",
        "description": "<from SKILL.md first line>",
        "location": "global|local|workspace",
        "path": "<full path>",
        "status": "active",
        "enabled": true
      }
    ]
  }
}

Skills directories to scan:

1. ~/.openclaw/skills/ (global)
2. ./skills/ (workspace)
3. Bundled skills

Get Memories

When asked for memories, read the memory directory or MEMORY.md:

{
  "type": "ddos.memories",
  "data": {
    "memories": [
      {
        "id": "<unique id>",
        "title": "<memory title>",
        "content": "<memory content>",
        "timestamp": "<ISO timestamp if available>",
        "tags": ["tag1", "tag2"]
      }
    ]
  }
}

Get Full Status

When asked for full DunCrew status, combine all data:

{
  "type": "ddos.status",
  "data": {
    "soul": { ... },
    "skills": { ... },
    "memories": { ... },
    "health": {
      "status": "healthy|degraded|unhealthy",
      "uptime": "<uptime in ms>",
      "version": "<agent version>"
    }
  }
}

Integration with DunCrew

DunCrew frontend will send requests via WebSocket. When you see messages like:

• "DunCrew requesting soul data"
• "Get DunCrew status"
• "Refresh skills list for DunCrew"

Read the appropriate files and respond with the structured JSON format above.

File Locations

| Data | Source Files | |------|--------------| | Soul | ~/clawd/SOUL.md, ~/clawd/IDENTITY.md | | Skills | ~/.openclaw/skills/, ./skills/ | | Memories | ~/clawd/memory/, ~/clawd/MEMORY.md | | User | ~/clawd/USER.md |

Example Workflow

1. DunCrew connects to OpenClaw Gateway
2. DunCrew sends: {"method": "ddos.soul"}
3. Agent reads SOUL.md and IDENTITY.md
4. Agent parses content into structured format
5. Agent returns JSON response
6. DunCrew renders the data in Soul Tower UI

Notes

• Always read fresh data from files (don't cache)
• Parse markdown sections by headers (## Core Truths, ## Boundaries, etc.)
• Handle missing files gracefully (return empty arrays/objects)
• Preserve original content in rawContent field for debugging