-
Notifications
You must be signed in to change notification settings - Fork 0
Home
Caduceus is a Hermes Agent plugin daemon that watches your watched GitHub repositories, claims issues, dispatches workers, and finalises the result — commit, push, pull request, comment — with durable checkpoints at every step. A crash mid-work resumes from the last checkpoint instead of re-running side effects.
The daemon is one Rust binary, one Python bridge, one Hermes plugin. It is small on purpose: the supervisor contract is the design, not a framework.
Never edit daemon state, claim files, or transcripts directly. State recovery has its own page — follow it when something is corrupt or wedged:
- State Recovery — corruption, migration, queue reset, lock cleanup
Pick the page that matches what you're trying to do:
- Installation — install on a new host (Hermes plugin path or standalone)
- Configuration — the full config schema
-
The Bridge — write or modify
worker-bridge.py - Architecture — internal design
- Plugin Lifecycle — the Hermes plugin lifecycle
- Hermes Integration — cron, gateway, slash commands
- CI — the GitHub Actions matrix and required-check contract
- Troubleshooting — something broke
- FAQ — common questions
- Public Voice — the public-voice rule
The migration procedure for replacing a prior install is in the repository README. Operators with an outage should not have to navigate a docs tree.
Contributor guidance lives in the repository: AGENTS.md, CONTRIBUTING.md, and the README.
Caduceus docs
- Home — what it is, and the one rule
- Installation — get it running
- Configuration — the settings that matter
- The-Bridge — make the worker yours
- State-Recovery — when things go wrong
- Troubleshooting — fix it
- FAQ — quick answers