Where does the bug appear (feature/product)?
Cursor IDE
Describe the Bug
Hello Cursor Support,
I am repeatedly encountering “Request blocked by Anthropic” while using Claude Opus 5.5 in Cursor Agent.
The task is ordinary software design work for a business SaaS application. It involves designing and planning a dictionary-management feature, including UI design, database/data design, and generating a static HTML mockup. It does not involve cybersecurity, penetration testing, exploitation, malware, credential access, or any other security-sensitive activity.
One of the Request IDs is:
Request ID: 37e1b323-ddc5-40eb-b5b3-5f5ae13d7ccb
{“error”:“ERROR_OPENAI”,“details”:{“title”:“Request blocked by Anthropic”,“detail”:“We are unable to complete this request because it was blocked under Anthropic’s Usage Policy.”,“isRetryable”:false,“additionalInfo”:{},“buttons”:,“planChoices”:,“analyticsMetadata”:{“actionRequired”:“cyber_policy_review”}},“isExpected”:true}
We are unable to complete this request because it was blocked under Anthropic’s Usage Policy.
Request blocked by Anthropic We are unable to complete this request because it was blocked under Anthropic’s Usage Policy.
ActionRequiredError: Request blocked by Anthropic We are unable to complete this request because it was blocked under Anthropic’s Usage Policy.
at kbS (vscode-file://vscode-app/Applications/Cursor.app/Contents/Resources/app/out/vs/workbench/workbench.glass.main.js:7536:18201)
at sxa (vscode-file://vscode-app/Applications/Cursor.app/Contents/Resources/app/out/vs/workbench/workbench.glass.main.js:7536:16853)
at XbS (vscode-file://vscode-app/Applications/Cursor.app/Contents/Resources/app/out/vs/workbench/workbench.glass.main.js:7537:13913)
at Object.classify (vscode-file://vscode-app/Applications/Cursor.app/Contents/Resources/app/out/vs/workbench/workbench.glass.main.js:7537:17483)
at GbS (vscode-file://vscode-app/Applications/Cursor.app/Contents/Resources/app/out/vs/workbench/workbench.glass.main.js:7537:6366)
at async gxa.run (vscode-file://vscode-app/Applications/Cursor.app/Contents/Resources/app/out/vs/workbench/workbench.glass.main.js:7537:16790)
at async Rve.runAgentLoop (vscode-file://vscode-app/Applications/Cursor.app/Contents/Resources/app/out/vs/workbench/workbench.glass.main.js:8939:31457)
at async Foc.streamFromAgentBackend (vscode-file://vscode-app/Applications/Cursor.app/Contents/Resources/app/out/vs/workbench/workbench.glass.main.js:20940:17007)
at async Foc.getAgentStreamResponse (vscode-file://vscode-app/Applications/Cursor.app/Contents/Resources/app/out/vs/workbench/workbench.glass.main.js:20940:39112)
at async cGe._submitChatMaybeAbortCurrent (vscode-file://vscode-app/Applications/Cursor.app/Contents/Resources/app/out/vs/workbench/workbench.glass.main.js:8917:4491)
I have retried the task several times, including:
- starting a new Agent chat,
- reducing/rewording the prompt,
- explicitly clarifying that the task is normal business application development.
However, the request continues to be blocked.
The original task asks Cursor to inspect an existing application and create an implementation/design plan for consolidating two existing terminology/pronunciation configuration features into a new “Dictionary Settings” feature, including UI and data design and a static HTML mockup.
Given the benign nature of the task and the repeated reproducibility, this appears to be a false positive in the Anthropic cyber-policy classifier rather than an actual Usage Policy violation.
Could you please investigate the Request ID(s), confirm what part of the context is triggering the policy block, and, if appropriate, escalate this to Anthropic as a false-positive classification?
I can also provide the full original prompt and additional Request IDs from the repeated attempts if needed.
Thank you.
Steps to Reproduce
Launch new session and send the task as follows:
このシステムで通話用AIに業務専門用語やその読みの情報を与える機能として、現状では以下の2つがあります。
- 「案件設定」の中の「業務用語設定」
- 「受電(架電)応対パターン設定」の中の「語句と読みの個別指定」
これらを独立した機能として切り出しつつ統合し、新たに「辞書設定」というメニューと画面を設けたいです。
その機能の要件は以下のとおりとしたいです。
- 「共通辞書」と「トークフロー別辞書」という2つの括りを設ける。
- 「共通辞書」側は、ユーザーが特段の紐付け設定をしなくても、そのユーザー企業が持つ全トークフローに適用され、通話用AIがそれを知っている前提で会話に使用される。
- 一方で、「トークフロー別辞書」側は、ユーザーが紐付けた受電(架電)応対パターンでのみ使用される。
- 「共通辞書」は1つの辞書に複数語句を持つ形とし、「トークフロー別辞書」はn個の辞書それぞれに複数語句を持つ形とする。
- どちらの辞書にも、CSVを介する形でのインポート機能とエクスポート機能を設ける（既存機能と同様のUXを踏襲）。
- 「トークフロー別辞書」は複数の受電(架電)応対パターンに紐付けることができる。
- 「トークフロー別辞書」と受電(架電)応対パターンとの紐付けは、「トークフロー別辞書」の管理画面からもでき、受電(架電)応対パターンの画面からもできるようにする。
- 受電(架電)応対パターン側は、既存の「語句と読みの個別指定」欄のあたりにその紐付け用UIを差し替えで設置する形を想定している。
- 「案件設定」の「業務用語設定」はそのタブごと廃止してしまう。
- 「辞書設定」画面の大枠のレイアウトは、現在の「提供AI管理」と同じようにしてしまう。
- 画面のほぼ全体を使い、そこに「共通辞書」と「トークフロー別辞書」という2つのタブを設ける。
- それぞれのタブの中身は、UI/UX的に適切と考えられるUIを構成する。
上記の「共通辞書」機能について、実装計画と設計（UI設計やデータ設計を含む）をし、その内容を todo 直下に新規のHTMLファイルとして書き出してください。
そのHTML内に、UI設計確認用の静的モックも載せてください。
Expected Behavior
WORK CORRECTLY.
Operating System
MacOS
Version Information
Version: 3.22.12
VS Code Extension API: 1.128.0
Commit: 3a92974361033b2051526321308c2740fe5912c0
Date: 2026-09-26T04:55:12.125Z
Layout: Agent Window
Build Type: Stable
Release Track: Default
Electron: 42.10.0
Chromium: 148.0.7778.280
Node.js: 24.18.1
V8: 14.8.178.38-electron.0
xterm.js: 6.1.0-beta.291
OS: Darwin arm64 25.6.0
For AI issues: which model did you use?
Claude Opus 5.5
For AI issues: add Request ID with privacy disabled
Request ID: 37e1b323-ddc5-40eb-b5b3-5f5ae13d7ccb
{“error”:“ERROR_OPENAI”,“details”:{“title”:“Request blocked by Anthropic”,“detail”:“We are unable to complete this request because it was blocked under Anthropic’s Usage Policy.”,“isRetryable”:false,“additionalInfo”:{},“buttons”:,“planChoices”:,“analyticsMetadata”:{“actionRequired”:“cyber_policy_review”}},“isExpected”:true}
We are unable to complete this request because it was blocked under Anthropic’s Usage Policy.
Request blocked by Anthropic We are unable to complete this request because it was blocked under Anthropic’s Usage Policy.
ActionRequiredError: Request blocked by Anthropic We are unable to complete this request because it was blocked under Anthropic’s Usage Policy.
at kbS (vscode-file://vscode-app/Applications/Cursor.app/Contents/Resources/app/out/vs/workbench/workbench.glass.main.js:7536:18201)
at sxa (vscode-file://vscode-app/Applications/Cursor.app/Contents/Resources/app/out/vs/workbench/workbench.glass.main.js:7536:16853)
at XbS (vscode-file://vscode-app/Applications/Cursor.app/Contents/Resources/app/out/vs/workbench/workbench.glass.main.js:7537:13913)
at Object.classify (vscode-file://vscode-app/Applications/Cursor.app/Contents/Resources/app/out/vs/workbench/workbench.glass.main.js:7537:17483)
at GbS (vscode-file://vscode-app/Applications/Cursor.app/Contents/Resources/app/out/vs/workbench/workbench.glass.main.js:7537:6366)
at async gxa.run (vscode-file://vscode-app/Applications/Cursor.app/Contents/Resources/app/out/vs/workbench/workbench.glass.main.js:7537:16790)
at async Rve.runAgentLoop (vscode-file://vscode-app/Applications/Cursor.app/Contents/Resources/app/out/vs/workbench/workbench.glass.main.js:8939:31457)
at async Foc.streamFromAgentBackend (vscode-file://vscode-app/Applications/Cursor.app/Contents/Resources/app/out/vs/workbench/workbench.glass.main.js:20940:17007)
at async Foc.getAgentStreamResponse (vscode-file://vscode-app/Applications/Cursor.app/Contents/Resources/app/out/vs/workbench/workbench.glass.main.js:20940:39112)
at async cGe._submitChatMaybeAbortCurrent (vscode-file://vscode-app/Applications/Cursor.app/Contents/Resources/app/out/vs/workbench/workbench.glass.main.js:8917:4491)
Does this stop you from using Cursor
Sometimes - I can sometimes use Cursor