SECURITY • OpenClaw quick fix
Featured: @AmpereSh
Exposed openclaw.json? Rotate before agents run.
A field report warned that exposed OpenClaw config files can become a model-credit-drain path. The exact incident still needs independent confirmation, but the risk is real: OpenClaw config can contain provider keys, tokens, app secrets, and gateway settings, and a closed GitHub issue showed how raw config exposure can leak plaintext credentials through tool surfaces.
In simple words: Make a safe copy if the step could change your setup, try the smallest check, then confirm OpenClaw is back to normal before doing more. The source link and commands stay below for people who want the technical detail.
Checklist
Keep `~/.openclaw`, `openclaw.json`, env files, logs, and backup archives out of public web roots, shared static folders, repos, and unauthenticated file browsers.
Fast local exposure sweep
# Run from any repo or deployed static root you control.
find . -maxdepth 4 -type f \( -name 'openclaw.json' -o -name '.env*' -o -name '*backup*.zip' \) -print
Check every self-hosted box, tunnel, and reverse proxy: only the intended gateway/control surfaces should be reachable, and admin/config paths should require trusted access.
If config may have leaked, rotate provider keys and bot/app secrets first; do not wait for proof that credits have already been used.
Set provider-side spend caps, alerts, and per-key separation so a leaked agent key cannot silently burn the whole account.
After rotating, run a small known-good agent task and confirm logs no longer reference the old key or exposed config path.
Success looks like
• No OpenClaw config, env, logs, or backup artifacts are reachable from public URLs or shared static roots.
• Any possibly exposed provider/bot credentials have been rotated and capped.
• The next agent run uses the new secret path and stays within provider-side budget alerts.
Source: GitHub issue + X field report · link