ClawHub off, third-party skills rewritten locally

Disable ClawHub and rewrite third-party skills locally. Security-minded OpenClaw runbook.

Share on 𝕏 ↗ Compare with… Read the source ↗

what to steal

Read and rewrite any third-party skill before it runs on your machine.

Our takeaway from the source, not a quote.

Drawn from this card's sourced fields. Amber = older model named by the source. Rose = cut or replaced.

digitalknk-openclaw-hardening.toml

# Disable ClawHub and rewrite third-party skills locally
# source: https://github.com/digitalknk/openclaw-runbook

[rig]
who     = "Security-minded OpenClaw runbook"
author  = "digitalknk (GitHub)"
date    = "not stated"

harness = ["OpenClaw"]
models  = []

[keep]
  - Gateway loopback-only / Tailscale
  - Keys in auth profiles, not openclaw.json
  - Narrow local skills
  - Cheaper explicit models for cron/heartbeat
[cut]
  - ClawHub
  - Public gateway exposure

thing	status	why (per source)
ClawHub	disabled by default	Inspect source and write a local skill instead
Public gateway exposure	removed	Use Tailscale or local-only

Source: https://github.com/digitalknk/openclaw-runbook · checked 2026-10-14

OpenClaw and Hermes in separate VMs, so neither is a single point of failure

steal → Put each agent in its own snapshot-able VM; fix credentials next.

OpenClawHermes AgentCodex GPT-5.5Qwen3.6 27B / 35B Credentials in the age

by Eli Lap

~/rigs/justinflick-openclaw-routing.toml2026-03-28

Cheap heartbeat model, API fallbacks, aliases for /model

steal → Run cron/heartbeat on a cheap model and keep the strong one for real work.

OpenClaw openai/gpt-5-4google/gemini-3-flash-

by Justin Flick

~/rigs/mrclaw207-openclaw-247.tomlundated

One OpenClaw agent 24/7 plus one research subagent

steal → Start with one main agent and one helper, not a swarm.

OpenClawClaude Code claude-sonnet-4-202505 Many agents up front

by mrclaw207 (DEV)