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

I have the same situation. The same bug.

Same issue, needs fix.

same issue...

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.

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.

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!

Same bug, opposite direction from the original report. My user account was reinstated but my org is still flagged.

Timeline:

• Had Max 20x, used Claude Code daily via OAuth login
• Account suspended, then Anthropic refunded the subscription and lifted the user-level suspension
• I can log in, see my chat history, use Free tier features
• But the org still has a banned flag, so I can't resubscribe or send messages

Subscription attempt:

PUT /api/organizations/[org_id]/subscription
HTTP 400

{"error":{"message":"Cannot start subscription for a banned organization","details":{"error_code":"account_banned"}}}

Sending a message on Free tier:

POST /api/organizations/[org_id]/chat_conversations/[conv_id]/completion
HTTP 400

{"error":{"message":"This organization has been disabled.","details":{"error_code":"organization_disabled"}}}

The org GET still returns 200 with api_disabled_reason: "trust_and_safety", but active_flags: [] and capabilities: ["chat"]. So read works, writes don't.

The appeal form at /restricted redirects to /new because my user layer is clean. usersafety@ sends bot replies. I have the org UUID and both request IDs if anyone from Anthropic wants to look into this.

Claude Code 2.1.197, macOS.

Hi,

Same bug here on a personal account Claude MAX:

• Account banned on Friday 17, 1.56 am UTC +4: I received 3 emails quoting different references informing me that my account was suspended (Reference TS-019f6cea-965d-72c7-a360-6fb85eec6776 - TS-019f6cea-9997-75aa-a102-2d5e98035ecd and TS-019f6cea-0e30-75ed-8ab1-c247bc3aba47)
• Appealed in the morning after waking up
• Received the same day a confirmation that my appeal was successful and my account reinstated (Reference TS-019f714f-1b5a-762a-80c4-6018071c50b5 )

48 hours later, I still can't access my account. After logging in, my account still appears banned despire the reinstatement, or it loops back to the "enter your email address login page".

I have written to support but just receive bot answer telling me the ticket is closed.

Let me know if you manage to recover your access and how to.

Thank you

Same bug, opposite direction from the original report. My user account was reinstated but my org is still flagged. Timeline: Had Max 20x, used Claude Code daily via OAuth login Account suspended, then Anthropic refunded the subscription and lifted the user-level suspension I can log in, see my chat history, use Free tier features But the org still has a banned flag, so I can't resubscribe or send messages Subscription attempt: `` PUT /api/organizations/[org_id]/subscription HTTP 400 {"error":{"message":"Cannot start subscription for a banned organization","details":{"error_code":"account_banned"}}} ` Sending a message on Free tier: ` POST /api/organizations/[org_id]/chat_conversations/[conv_id]/completion HTTP 400 {"error":{"message":"This organization has been disabled.","details":{"error_code":"organization_disabled"}}} ` The org GET still returns 200 with api_disabled_reason: "trust_and_safety", but active_flags: [] and capabilities: ["chat"]`. So read works, writes don't. The appeal form at /restricted redirects to /new because my user layer is clean. usersafety@ sends bot replies. I have the org UUID and both request IDs if anyone from Anthropic wants to look into this. Claude Code 2.1.197, macOS.

I received another 2 emails today after nearly 3 weeks stating that my account has been banned and I can finally submit my appeal form through claude.ai/restricted. The problem solved but the issue remains stale for many other user.

Any of you managed to get in touch with a human at support? I am stucked in an endless loop where nobody cares.

Has anyone found a reliable solution for this? I’m experiencing the same issue on my personal account. Account suspended then reinstated within 24 hours yet logging in shows I’m still banned. Is there any way to reach human support? Is there any solution people have found or is it time to create a new account?