OpenClaw dm.open Finding (channels.<provider>.dm.open): What It Means and How to Fix It

Published 2026-10-20 · Updated 2026-10-20 · By The Is My Agent Safe team

Short answer

A chat channel's DM policy is open, so anyone who can message the bot can give your agent instructions. Fix it first: stop the Gateway, set that channel's dmPolicy to pairing (or allowlist), start the Gateway and run openclaw security audit --deep again. Then approve the people you want with openclaw pairing approve.

  • OpenClaw rates channels.<provider>.dm.open critical. The <provider> part is the channel, such as telegram or discord.
  • Quickest fix: set the channel's dmPolicy to pairing, so unknown senders get a code and their messages aren't processed until you approve.
  • dm.open_invalid (warn) means dmPolicy is open but allowFrom has no "*". OpenClaw's docs say only the listed senders get in, so allowlist is the honest setting.
  • An open GitHub report (#115410) says the audit doesn't check DM policy on Microsoft Teams or Feishu. OpenClaw's maintainers haven't confirmed it or shipped a fix.

Paste your audit: run openclaw security audit --deep and paste the output into our free check for a red/amber/green result with the most important finding first (Limits of the check).

What should I do first?

Switch the channel named in the finding to pairing:

openclaw gateway stop
openclaw config set channels.<channel>.dmPolicy pairing   # e.g. channels.telegram.dmPolicy
openclaw config validate
openclaw gateway start
openclaw security audit --deep

The finding's detail names the exact setting. Made-up example:

channels.telegram.dmPolicy="open" allows anyone to DM the bot.

If it names an account, such as channels.telegram.accounts.work.dmPolicy, use that path. OpenClaw's config docs say config set needs the Gateway stopped, which is why the commands stop it first. When the audit comes back without dm.open, you're done with this finding.

Then let in the people you want. OpenClaw's pairing docs say that with pairing, "unknown senders get a short code and their message is not processed until you approve". Approve them from the Control UI (Settings → Channels → DM access requests) or the CLI:

openclaw pairing list telegram
openclaw pairing approve telegram <CODE>

The same docs say codes "expire after 1 hour" and pending requests are capped at 3 per channel account. Approval "grants direct-message access only. It does not grant group access."

Pairing works on channels that implement OpenClaw's pairing API. On other channels, set dmPolicy to allowlist and list the sender IDs you trust in allowFrom.

What does channels.<provider>.dm.open mean?

OpenClaw's security audit checks page lists it as critical: "<provider> DM policy is "open"; anyone can DM the bot". The settings it points to are the channel's dmPolicy and allowFrom.

In OpenClaw 2026.10.5's source, the finding appears for each channel account whose DM policy is open. OpenClaw's own suggested fix, with your channel's path filled in, is:

“Use pairing/allowlist; if you really need open DMs, ensure channels.<channel>.allowFrom includes "*".”

What does dm.open_invalid mean?

channels.<provider>.dm.open_invalid is a warn that appears alongside dm.open when allowFrom doesn't include "*". Made-up example of its detail:

"open" requires channels.telegram.allowFrom to include "*".

OpenClaw's pairing docs explain what that state does: "If existing state contains open with concrete allowFrom entries, runtime still admits only those senders, and pairing-store approvals do not widen open access." So the channel already works like an allowlist. Make the config say so:

openclaw gateway stop
openclaw config set channels.<channel>.dmPolicy allowlist
openclaw config validate
openclaw gateway start
openclaw security audit --deep

Why does OpenClaw rate it critical?

Anyone who can message the bot can give your agent instructions. OpenClaw's audit triage order puts this first:

“Anything "open" + tools enabled: lock down DMs/groups first (pairing/allowlists), then tighten tool policy/sandboxing.”

If tools are reachable from that channel, you'll often see security.exposure.* findings at the same time. Closing the DM policy is the first step for those too. See every audit finding, explained.

If you want public DMs on purpose

Some agents are meant to talk to anyone. In that case, keep "*" in allowFrom (OpenClaw's docs say open "is public only when the effective DM allowlist includes "*""), keep that agent's tools to a minimum and expect this finding to stay. OpenClaw still rates it critical.

A maintainer closed issue #154931 as fixed by pull request #153147. The comment says that during an update, Doctor's lint now treats intentional open-DM findings as warnings, so they no longer stop openclaw update. It also says "standalone lint/audit severity is unchanged". GitHub shows that change is included in 2026.10.5.

What if a stranger already messaged the agent?

The finding describes your config, not an event. If people you don't know could message the agent, follow OpenClaw's operator incident response plan in full. It covers containing access, rotating Gateway and provider credentials, reviewing logs, transcripts and recent config changes, and re-running openclaw security audit --deep.

Could it be a false alarm?

Confirm what the config says for that channel:

openclaw config get channels.<channel>

If dmPolicy really is "open", the finding matches your config.

The open report on this checkId is about the opposite problem. Issue #115410 says the audit only checks DM policy on channels whose plugin supports that check, and names Microsoft Teams and Feishu as channels where dmPolicy: "open" produces no finding. OpenClaw's maintainers haven't confirmed it or shipped a fix. So a missing dm.open finding doesn't prove that every channel's DMs are closed. Check each channel's dmPolicy yourself.

For findings you think are wrong, our guide to audit false alarms shows how to check a finding and report it. Don't suppress a critical finding you haven't confirmed.

What our free check covers

The free check at /check is paste-only. We check the text you paste: your openclaw security audit output. The check never connects to your Gateway or your chat channels, so it can't see who has messaged your agent or check channels the audit didn't report. Nothing is installed into the agent you're checking. Green means no issues found in the checks we ran.

The free check puts the most important finding first, in plain English. The full report explains every finding and lists the fixes in order. The check doesn't detect malicious skills. We don't review the code inside skills or MCP servers ourselves, beyond what OpenClaw's own audit reports. The check is free, and a full report is $29. See Limits of the check.

Frequently asked questions

What does channels.<provider>.dm.open mean?

OpenClaw's audit found a channel whose DM policy is open. Its detail says the setting "allows anyone to DM the bot". OpenClaw rates it critical. The <provider> part names the channel, for example channels.telegram.dm.open.

How do I fix channels.<provider>.dm.open?

Stop the Gateway, run openclaw config set channels.<channel>.dmPolicy pairing (use the exact path from the finding's detail), validate the config, start the Gateway and re-run openclaw security audit --deep. Then approve the senders you want with openclaw pairing approve <channel> <CODE>.

What does dm.open_invalid mean?

It's a warning that appears with dm.open when the DM policy is open but the channel's allowFrom list doesn't include "*". OpenClaw's pairing docs say that, in this state, only the listed senders are let in. Setting dmPolicy to allowlist matches what the channel already does.

Does this finding mean a stranger used my agent?

No. It describes your config, not an event. If people you don't know could message the agent, follow OpenClaw's operator incident response plan in full. It covers closing access, rotating Gateway and provider credentials, reviewing logs, transcripts and config changes, and re-running the audit.

Can your free check see my chat channels?

No. We check the text you paste, which is where this finding comes from. The check never connects to your Gateway or your chat channels, so it can't see who has messaged your agent.

Sources

  1. OpenClaw docs: security audit checks (checkId catalog)
  2. OpenClaw docs: pairing (DM access approval)
  3. OpenClaw docs: openclaw pairing (CLI)
  4. OpenClaw docs: running the security audit (triage order)
  5. OpenClaw docs: operator incident response
  6. OpenClaw docs: openclaw config (CLI)
  7. OpenClaw source v2026.10.5: src/security/audit-channel.ts (where the check is defined)
  8. GitHub issue #115410 (open): security audit skips DM policy for channels without resolveDmPolicy
  9. GitHub issue #154931 (closed): update path for an intentionally public DM agent