AI Agent Hub
Back to skills
Empire Genesis icon

Empire Genesis

AI Agent Updated 2026.08.30

Paste the following prompt into your AI chat to install this skill:

Please install @user_f9be8d90/empire-genesis into your AI assistant following https://skillhub.cn/install/skillhub.md.

About this skill

Problem

When multi-agent systems are deployed across operating systems, state storage, timestamps, and heartbeats become failure points: SQLite can hit concurrent locking on Windows, duplicate alerts amplify noise, and mixing event_time with collection time across agents makes debugging harder.

How It Works

Empire Genesis bootstraps OpenClaw into a self-checking system architecture. Key steps include:
- Creating a concurrency-safe StateDB, using lesson_hash as a unique index and ON CONFLICT to increment hit_count, which suppresses alert storms.
- Injecting Signal Protocol v2 and a dual-timestamp mechanism, preserving both event_time and collection_time in the blackboard table.
- Adding Research, Execution, and Verification roles, then deploying a Watchdog sentinel that periodically refreshes last_seen for online static agents.
- Running sandbox tests to validate StateDB throughput and constraint logic, then self-deleting after a passing result.

Boundaries

The skill is intended for explicit commands such as initializing the empire or deploying the lobster architecture, and is focused on low-level assembly and self-checks. It is not suitable for invoking a single agent temporarily. Known behavior includes Watchdog leaving main and its own heartbeat untouched, but before rollout you should confirm OS behavior, load, and constraints such as timeout=20.

Use Cases

  • Bootstrap OpenClaw state storage on Windows and avoid SQLite concurrent Database is locked errors.
  • Deploy a multi-agent foundation with consistent event_time and collection_time for cross-agent debugging.
  • Create StateDB for Research, Execution, and Verification roles, then configure Watchdog to refresh online heartbeats.
  • Run sandbox checks on StateDB throughput and alert de-duplication, then delete test artifacts after passing.

Best For

  • Engineers deploying OpenClaw multi-agent systems who need a self-checking architecture bootstrap.
  • Ops engineers maintaining agent heartbeats who want Watchdog to reduce online static agent false negatives.
  • Backend engineers designing alert de-duplication who need lesson_hash indexing to suppress duplicate alert storms.
  • Platform engineers doing pre-release validation who require StateDB to pass sandbox throughput and constraint checks before deletion.