| name | openclaw-hermes-upgrade-kit |
| description | Install two Hermes-like capabilities into OpenClaw: a learning loop for self-improvement and Honcho-based user modeling for richer cross-session memory. Use when the user wants a packaged OpenClaw upgrade kit, a public tutorial bundle, or a one-click distribution package that combines OpenClaw Learning Loop and Honcho onboarding. |
OpenClaw Hermes Upgrade Kit
Use this skill when the user wants to package or install a combined upgrade bundle for OpenClaw that includes:
- a learning loop / self-improvement workflow
- Honcho-based user modeling and shared memory
This skill is for:
- public-facing delivery bundles
- installable OpenClaw add-on kits
- “Hermes-like experience” packaging for tutorials
- combining multiple reusable sub-skills into one distribution
This skill is not for:
- claiming feature parity with Hermes
- publishing private paths, keys, or identities
- promising fully automatic orchestration
Package Goal
The package should help a user add two missing layers to OpenClaw:
-
Learning Loop
Turn solved problems into reusable learning artifacts and future skills.
-
Honcho User Modeling
Add user profile inference, cross-session recall, and better multi-agent context sharing.
Delivery Shape
Build the package as:
- one top-level installable kit
- two bundled sub-skills
- one-click install script
- local Honcho deployment scaffold
- verification instructions
Safe Publishing Rules
- remove all real API keys
- remove personal workspace names
- remove personal filesystem paths
- use neutral peer names and workspace labels
- do not embed private logs or personal memory content
Recommended Bundle Layout
openclaw-hermes-upgrade-kit/
├── SKILL.md
├── README.md
├── INSTALL.md
├── run.command
├── install-openclaw-hermes-upgrade.sh
├── scripts/
│ └── verify_install.sh
├── templates/
│ ├── honcho.env.example
│ └── HONCHO_NEXT_STEPS.md
└── bundled-skills/
├── openclaw-learning-loop/
└── honcho-memory-starter/
Operating Rules
- Keep the learning-loop skill self-contained and path-agnostic.
- Keep the Honcho skill focused on memory modeling, not orchestration.
- Make the installer copy bundled skills into the user's OpenClaw workspace.
- If cloning Honcho, clone the public upstream repository rather than shipping a private fork.
- Scaffold configuration safely with
.env.example, never real secrets.
Output Contract
When building this package, provide:
- a reusable top-level skill
- a public README
- a clear installation guide
- a one-click installer
- bundled sub-skills
- verification guidance
