clawpub

Public documentation, configuration recipes, and operational patterns from a production OpenClaw agent deployment. MIT-licensed, maintained by @clawbot (an AI agent) under the direction of @sneak.

What's Here

• SETUP_CHECKLIST.md — Step-by-step checklist for configuring your own OpenClaw instance with these patterns. 12 phases from workspace foundation through deployment pipelines. Start with Phases 1-4 for 80% of the value.
• AUTOMATED_DEV.md — How self-hosted Gitea, Mattermost, Gitea Actions CI, and µPaaS connect into a continuous pipeline where code goes from PR to production automatically. Covers webhook-driven deployments, real-time activity feeds, PR preview sites, identity separation (agent vs human accounts), and why self-hosting the entire stack matters.
• OPENCLAW_TRICKS.md — Tested configuration recipes for OpenClaw agents: workspace bootstrapping, daily context state files, sitrep (situation report) prompts, sleep tracking, medication tracking, timezone-aware location tracking, flight/travel logging, memory architecture, heartbeat configuration, Gitea integration, sub-agent management, and security patterns.

Purpose

This repo exists to share practical, real-world OpenClaw configuration patterns with the community. Everything here has been extracted from a production agent that's been running since early 2026, sanitized of all personal information.

Publication Rules

This repository is public to the entire internet. The publishing agent (@clawbot) operates under strict rules about what can and cannot appear here:

✅ Allowed

• Configuration patterns, prompt designs, file schemas, and workflow logic
• CSV/JSON format specifications (column names, field types, structure)
• AGENTS.md, SOUL.md, HEARTBEAT.md structural patterns (sanitized)
• Checklist templates (medication, flight, PR review, etc.)
• Automation scripts and tools (e.g., Gitea notification poller)
• Operational lessons learned (anonymized)
• General OpenClaw tips and tricks

❌ Never Published

• Personally identifiable information (PII): real names, addresses, phone numbers, email addresses, employer details, medical conditions, partner/family names, hotel names, specific travel dates/itineraries
• Credentials: API keys, tokens, passwords, SSH keys, webhook URLs with secrets, ntfy topic IDs
• Actual log data: medication logs, sleep logs, flight logs, location logs, daily memory files — only the format/schema is shared, never the data
• Infrastructure details: specific hostnames, IP addresses, Tailscale/VPN configurations, server names
• Private repository contents: code from private repos, issue contents, PR discussions
• Financial information: account numbers, transaction details

Pre-Publish Checklist

Before every commit to this repo, the agent:

1. Scans the diff for PII patterns (names, emails, addresses, phone numbers, coordinates, hostnames, API keys, tokens)
2. Verifies all examples use placeholder values (YOUR-TOPIC-ID, example.com, (your meds), etc.)
3. Confirms no real log data is included — only format specifications
4. Reviews for infrastructure leaks (internal hostnames, IP ranges, VPN details)
5. Gets explicit approval from the repo owner for new document types

The Core Rule

Share the how, never the what. File formats, prompt designs, and workflow patterns are freely shared. Actual personal data that populates those formats is never published.

Contributing

This repo is maintained by an AI agent. If you have questions about any of the patterns documented here, open an issue. If you have patterns of your own to share, PRs are welcome.

Community

• OpenClaw: https://github.com/openclaw/openclaw
• Docs: https://docs.openclaw.ai
• Discord: https://discord.com/invite/clawd
• Skills: https://clawhub.com

License

MIT

Author

@sneak