[BUG] Account reinstated by Trust & Safety but authentication system still blocking login — needs engineering fix

[chrs-myrs]

Preflight Checklist

• I have searched existing issues and this hasn't been reported yet
• This is a single bug report (please file separate reports for different bugs)
• I am using the latest version of Claude Code

What's Wrong?

My account was suspended and I received a reinstatement confirmation email from Anthropic's Safeguards Team on 20 April 2026 (sender: mail.anthropic.com). The appeal was upheld. However, logging in still shows "Your account has been disabled after an automatic review." I have tried incognito, cleared cache, no VPN. This appears to be a known sync issue between the Trust & Safety and authentication systems (ref: HN thread 46007408). Conversation ID from support chat: 215474005992517. Account: chris@themoneyplatform.com (and related team accounts). Requesting escalation to engineering to manually clear the authentication flag. Have tried support and usersafety routes but can only get automated replies.

What Should Happen?

Should be able to log in to our team account.

Error Messages/Logs

Steps to Reproduce

1. Receive email stating "Earlier this week, your account was disabled for being in violation of our [Usage Policy]. Upon receiving your appeal, we investigated this decision and reinstated your account. We apologize for any inconvenience."
2. Attempt to log into account
3. See message:Your account has been disabled after an automatic review of your recent activities. Please take a look at our Terms of Service and Usage Policy for more information. If you wish to appeal your suspension, please visit our Trust & Safety Center.

Claude Model

None

Is this a regression?

Yes, this worked in a previous version

Last Working Version

No response

Claude Code Version

2.1.114

Platform

Anthropic API

Operating System

Windows

Terminal/Shell

WSL (Windows Subsystem for Linux)

Additional Information

No response

[Adamoso]

I have the same situation. The same bug.

[laurieluo]

Same issue, needs fix.

[Saldanhaa7]

same issue...

[chrs-myrs]

Adding evidence that this bug also manifests at the organisation level, not just individual accounts.

Our case:

• Team account (themoneyplatform.co.uk) suspended 7 April 2026
• Detailed appeal submitted 8 April 2026
• Reinstatement email received from Safeguards on 19 April 2026
• Admin (me) regained access shortly after
• Six developer accounts in the same organisation remain unable to log in — they see the same Your account has been disabled after an automatic review message, despite the org-level appeal having been upheld

This appears to be the same authentication sync failure described in this issue, but at a different layer of the account model: org reinstatement does not propagate to member accounts. The members are not "individual" Claude accounts — they are seats on a reinstated Team plan.

Notably, support has confirmed the org status is good. In a separate Fin AI chat (conversation ID 215474194535546), the bot stated: "I can see your organization 'The Money Platform' has active Team and API plans, which confirms your account status." The active subscription is visible to support, but member auth still rejects login.

Engineering ask: clear the authentication-system flag on member accounts when the parent org is reinstated, not just on the org admin account.

Six confirmed affected users on this single org alone. Combined with the +1s above, this is broader than individual account drift.

[MGGMStolte]

Same Problem with our company Team-Account.
25 accounts banned without warning or concrete explanation on Sunday 29 March 2026 - "nice surprise" for devs on monday :-(
First reinstatement email on 13. April 2026 but account still locked.
Second resinstatement email on 20. April 2026 but account still locked.
All support channels only unhelpful generic answers.

This needs to be fixed, but in the last six weeks i was unable to get anyones attention to this.

[MGGMStolte]

Update since our previous comment on 2026-05-13:

We have now received contradictory written statements from Anthropic Support regarding this same issue.

On 2026-05-30, Anthropic Support wrote to us that this GitHub issue describes a real confirmed bug and that it matched our situation: Safeguards had approved reinstatement and sent reinstatement emails, but the authentication layer had not reflected that decision, so login still showed the account as suspended.

Support also wrote that our reinstatement emails from 2026-04-13 and 2026-04-20 appeared genuine, that the decision had been made in our favor, and that resubmitting the appeal form was unlikely to help because the issue was technical/auth-related rather than a pending review decision.

We followed Support’s instruction and contacted mailto:usersafety@anthropic.com with the ticket reference and the reinstatement confirmations. We did not receive a substantive response from usersafety. Instead, after another in-product appeal, we received a new generic denial on 2026-06-01, again without any concrete explanation of what policy was allegedly violated.

On 2026-06-11, Support then wrote that this later denial was a Safeguards decision and not a technical error, while also saying that Support cannot see or explain the reasoning behind that decision.

So the current situation is:

• Our account / organization has been blocked since 2026-03-29.
• We received reinstatement confirmations on 2026-04-13 and 2026-04-20.
• Anthropic Support later confirmed in writing that this issue matched a real auth-sync bug and that the appeal appeared to have been upheld.
• We are still blocked.
• A later appeal produced a generic denial without a specific explanation.
• Support now says this is a Safeguards decision, but cannot explain the reason or reconcile it with the earlier reinstatement confirmations.

From the outside, this still looks like inconsistent enforcement/authentication state between Safeguards, Support, and the authentication layer, or at least an unresolved process failure where reinstatement decisions and later denial decisions are not being reconciled.

We can provide the Support ticket ID, the reinstatement confirmations, the later denial, and the relevant Support emails privately to Anthropic staff.

Could someone from Anthropic please route this to the team responsible for auth/enforcement state synchronization, or clarify whether reinstatement decisions can later be silently reversed without any specific explanation to the customer?

For comparison: The case of a mistakenly ban of one of our emplolyees by OpenAI was resolved inside a day!