🧑‍🚀 OpenClaw Spacesuit

Your AI agent's workspace, suited up for the real world.

A batteries-included framework scaffold for OpenClaw workspaces. When you install OpenClaw, you get an AI agent — but an agent without a workspace is like an astronaut without a spacesuit. This package gives your agent structure, memory, safety, and operational conventions.

What's Included

• Session startup protocol — security-first file loading order
• Memory system — daily logs + curated long-term memory with commit discipline
• Git workflow — mandatory pre-commit checks, worktree conventions, parallel multi-agent coordination with merge locks
• Safety rules — pre-action checklist, destructive-action guards, stand-down protocol
• Priority system — P0–P5 triage for tasks and incidents
• Cross-platform handoffs — structured context transfer between Slack/Discord/Telegram
• Heartbeat framework — proactive periodic checks with state tracking
• Decision logging — mandatory audit trail for architectural decisions
• Meta-learning framework — expert-first research methodology (Dunning-Kruger aware)
• Security baseline — secret transmission policy, prompt injection defense, data classification

Installation

Via ClawHub (recommended)

clawhub install spacesuit

This places the package at <workspace>/skills/spacesuit/.

Manual

cd <your-openclaw-workspace>
git clone https://github.com/jontsai/openclaw-spacesuit.git skills/spacesuit

First-Time Setup

After installation, run the installer to create workspace files from templates:

cd skills/spacesuit
make init

This creates AGENTS.md, SOUL.md, TOOLS.md, HEARTBEAT.md, SECURITY.md, MEMORY.md, IDENTITY.md, USER.md, and Makefile at your workspace root — each with framework content wrapped in SPACESUIT markers and space for your customizations.

Personalize

# Edit these files at your workspace root:
vi IDENTITY.md    # Name your AI
vi USER.md        # Tell it about yourself
vi SOUL.md        # Set the personality/vibe
vi TOOLS.md       # Add your tool configs

How It Works

OpenClaw reads hardcoded filenames from the workspace root (AGENTS.md, SOUL.md, etc.). Since we can't change that loading behavior, Spacesuit uses section-based merging:

1. Framework content is wrapped in <!-- SPACESUIT:BEGIN --> / <!-- SPACESUIT:END --> markers
2. On upgrade, only the content between markers is replaced
3. Everything outside the markers (your customizations) is preserved

Upgrading

cd skills/spacesuit
 
# See what would change
make diff
 
# Preview upgrade (no changes)
make upgrade-dry
 
# Apply upgrade
make upgrade
 
# Check version
make version

Package Structure

openclaw-spacesuit/
├── SKILL.md               # ClawHub metadata
├── VERSION                # Package version
├── CHANGELOG.md           # Release history
├── Makefile               # init/upgrade/diff commands
├── base/                  # Framework source content
│   ├── AGENTS.md          # Session protocol, memory, git, safety, priorities
│   ├── HEARTBEAT.md       # Heartbeat check skeleton
│   ├── MEMORY.md          # Long-term memory starter
│   ├── SECURITY.md        # Immutable security rules
│   ├── SOUL.md            # Personality scaffold
│   └── TOOLS.md           # Tool organization guide
├── templates/             # Workspace file templates
│   ├── AGENTS.md          # Framework + customization sections
│   ├── HEARTBEAT.md
│   ├── IDENTITY.md        # Agent identity (name, emoji)
│   ├── Makefile            # Gateway management
│   ├── MEMORY.md
│   ├── SECURITY.md
│   ├── SOUL.md
│   ├── TOOLS.md
│   └── USER.md            # About your human
└── scripts/               # Automation
    ├── install.sh          # First-time setup
    ├── upgrade.sh          # Section-based merge upgrade
    ├── diff.sh             # Show pending changes
    └── sync-operators.sh   # Generate operators.json from sessions

Data Layer Scripts

Spacesuit includes utility scripts that gather data for dashboards and tooling:

sync-operators.sh

Auto-generates state/operators.json from session transcripts:

# Run from workspace
./scripts/sync-operators.sh
 
# Preview without writing
./scripts/sync-operators.sh --dry-run
 
# Explicit workspace
./scripts/sync-operators.sh --workspace /path/to/workspace
 
# Multi-profile support (uses ~/.openclaw-<name>)
./scripts/sync-operators.sh --profile myprofile
./scripts/sync-operators.sh --dev  # shortcut for --profile dev
 
# Or via environment variable
OPENCLAW_PROFILE=myprofile ./scripts/sync-operators.sh

What it does:

• Scans OpenClaw session transcripts for user messages
• Extracts Slack user IDs and usernames
• Counts messages per operator
• Preserves manually-set roles across syncs

Output: state/operators.json — used by OpenClaw Command Center

Supported channels:

• ✅ Slack (] username (USERID): pattern)
• 🔜 Telegram, Discord, Signal (PRs welcome!)

Files Managed

Author

Created by jontsai.

License

MIT