RAP Sheet home
Install the bug reporting skill
Its theme's rank, 14 days: 2 on the Claude Code sheet
134 people reported

All time, since Sep 16, including 17 thumbs-up on GitHub.

Reasoning_extraction safeguard block

“Claude Code can't respond to this message with Opus 5.5.”

Is this your bug? What gets counted gets fixed. Pile on.

Paste this into your agent. It files your report, with your error and version.

I hit this bug: rapsheet.dev/sheets/claude-code/issues/3330/file
Follow it to file my report. Show me the text before you send.

No install. No account. Public and anonymous. The link asks your agent for one report, nothing else. What it sends

Fading. 11 reports and thumbs-up in the last 7 days and 27 the week before, down from 28 a week at the late-September peak. Not fixed, but a workaround is known.

Safeguards block messages with reasoning_extraction API error. Claude Code says it "can't respond to this message with Opus 5.5" during "safe, normal conversations", showing Details: [reasoning_extraction].

Is this your bug?

File my report
Where
API client
Error text
API Error: Opus 5.5's safeguards flagged this message (https://www.anthropic.com/legal/aup). This sometimes happens with safe, normal conversations. Claude Code can't respond to this message with Opus 5.5.Pasted in 8 reports
API Error: Fable 5.1's safeguards flagged this message (https://www.anthropic.com/legal/aup). This sometimes happens with safe, normal conversations. Claude…Pasted in 7 reports
API Error: Sonnet 5.5's safeguards flagged this message (https://www.anthropic.com/legal/aup). This sometimes happens with safe, normal conversations. Claude…Pasted in 7 reports
Versions
2.1.251 to 2.1.319. Still reported on the newest version, 2.1.319.
Platforms
Linux, macOS, Windows
How bad
Breaks a normal workflow.

Part of: Safeguards block session requests34 bugs

  1. Cyber safeguard blocks sessions377 people
  2. Cyber safeguard pauses sessions52 people
  3. Safeguards refuse work requests48 people
  4. Safeguards block requests47 people
  5. Classifier halts routine replies12 people
  6. Requests blocked with a [cyber] flag8 people
  7. Refusal on own-code hardening11 people
  8. SSH to own devices refused7 people
  9. Pentest work blocked6 people
  10. CTF requests flagged as malicious6 people
Show all 33Show only the largest 10
  1. Classifier blocks own-project work6 people
  2. Classifier halts reverse engineering5 people
  3. Safeguards flag verified cyber work6 people
  4. Firmware requests flagged as cyber6 people
  5. Classifier stops benign replies5 people
  6. reasoning_extraction API error5 people
  7. Safeguard blocks all messages4 people
  8. [bio] classifier blocks responses4 people
  9. Reasoning extraction safeguard block4 people
  10. Security investigation requests blocked3 people
  11. Classifier stops DeFi bot work3 people
  12. Safeguards block CVP members3 people
  13. "[cyber]" flag stops sessions1 person
  14. Safety classifier blocks agent3 people
  15. Fictional hacking content flagged3 people
  16. Cyber verification prompt recurs2 people
  17. Cyber safeguards block own devices2 people
  18. Safety classifier stops auth work2 people
  19. Phishing research project flagged2 people
  20. Cyber stop on report tasks2 people
  21. Cyber safety block on audits2 people
  22. Biological safeguard triggers in sessions with no biology content2 people
  23. Unexpected safeguards notice2 people

Original reports133, newest first

  1. Repeated "safeguards flagged this message" false positives (reasoning_extraction)#108093
  2. False positive AUP safeguard flag (reasoning_extraction) on benign debugging request#108073
  3. Safeguard Error: reasoning_extraction blocks normal work requests#107470
  4. Safeguards False Positive: reasoning_extraction Blocks Read-Only Subagent Review#107150
  5. Anthropic API Error: Opus 5.5 Safeguards Flag Normal Conversation#107137
  6. API Error: Opus 5.5 safeguards flag benign message (reasoning_extraction)#107041
  7. Opus 5.5 safeguards block reasoning trace request during race condition debugging#106997
  8. Safeguards Flag "reasoning_extraction" on Benign Pairing/Explanation Requests in Opus 5.5#106799
  9. Opus 5.5 safeguards flag benign message with reasoning_extraction error#106699
  10. reasoning_extraction error blocks simple one-line requests without tools#106683
Show the newest 100 of 133Show only the newest 10
  1. API Error: Opus 5.5 safeguards flagged message (reasoning_extraction)#106210
  2. Anthropic API Error: Content filtered by Claude Sonnet 5.5 safeguards#106115
  3. API Error: Sonnet 5.5 Safeguards Flagged Request (reasoning_extraction)#106048
  4. Claude Code refuses unrelated requests after asking about reasoning behind a change#106007
  5. API Error: Opus 5 Safeguards Flagged Message (reasoning_extraction)#105848
  6. I don't see any flag or "reasoning_extraction" label in this conversation, and I'm not able to see or change how messages get classified elsewhere. I can only see your message and the instructions I w#105762
  7. Opus 5.5 safeguards falsely flag benign messages with reasoning_extraction error#105740
  8. Reasoning extraction safeguard falsely flagged benign image-to-ASCII art question#105714
  9. API Error: Sonnet 5.5 Safeguards Flag Benign Message (reasoning_extraction)#105458
  10. Request blocked during document review validation#105407
  11. Bug Report: Unclear issue, no details provided#105359
  12. API Error: reasoning_extraction blocks request to document design rationale#105264
  13. Opus 4.5 safeguards block /compact command#105197
  14. API Error: Opus 5.5 Safeguards Flagged Message (reasoning_extraction)#105151
  15. Opus 5.5 safeguards falsely flag benign message "Can you read this message?"#105121
  16. Opus 5.5 Safeguards Flag Normal Message: reasoning_extraction#105051
  17. Anthropic API Error: Opus 5.5 Safeguards Flagged Normal Message (reasoning_extraction)#105019
  18. reasoning_extraction false positive (8/8): interrupted tool call on benign coding output — Opus 5.5 1M, Bedrock#105010
  19. reasoning_extraction false positive (7/8): interrupted tool call, then hard refusal on benign coding output — Opus 5.5 1M, Bedrock#105009
  20. reasoning_extraction false positive (6/8): interrupted tool call, then hard refusal on benign coding output — Opus 5.5 1M, Bedrock#105008
  21. reasoning_extraction false positive (5/8): interrupted tool call on benign coding output — Opus 5.5 1M, Bedrock#105007
  22. reasoning_extraction false positive (4/8): interrupted tool call on benign coding output — Opus 5.5 1M, Bedrock#105006
  23. reasoning_extraction false positive (3/8): hard refusal on benign coding output — Opus 5.5 1M, Bedrock#105005
  24. reasoning_extraction false positive (2/8): hard refusal on benign coding output — Opus 5.5 1M, Bedrock#105004
  25. reasoning_extraction false positive (1/8): hard refusal on benign coding output — Opus 5.5 1M, Bedrock#105003
  26. reasoning_extraction false positives: 8 incidents / 10 request IDs in one coding session (Opus 5.5 1M, Bedrock)#105002
  27. API Error: Safeguards Flagged Benign Message (reasoning_extraction)#104987
  28. False Positive Safeguard Flag on Git Commit Split Request#104817
  29. Sonnet 5.5 safeguards flag benign "save context and restart" request with reasoning_extraction error#104756
  30. Repeated safeguard blocks flagged as [reasoning_extraction]#104753
  31. reasoning_extraction classifier blocks a single-session token-cost question, even a plain "ok" reply#104746
  32. Opus 5.5 Safeguards Block Benign Message with reasoning_extraction Error#104722
  33. Output Style Flagged as [reasoning_extraction], Blocking Opus and Sonnet Access#104682
  34. Safety Classifier False Positives Cutting Off Routine Coding Replies (reasoning_extraction)#104653
  35. Safety classifier false positives repeatedly cut off routine coding replies (Opus 5.5)#104652
  36. Opus 5.5 safeguards block "reasoning_extraction" on benign project messages#104532
  37. Opus 5.5 safeguards falsely flag reasoning_extraction on normal conversations#104490
  38. False Positive reasoning_extraction Safeguard Flag on Tool Output#104373
  39. False Positive: reasoning_extraction Safeguard Blocks Normal Coding Requests on Opus 5.5#104364
  40. API Error: Opus 5.5 Safeguards Flag Benign Math Conversatio#104221
  41. False Positive: Legitimate PR Review Rationale Blocked as reasoning_extraction#104041
  42. Sessions flagged for reasoning extraction during normal Minecraft mod development#103852
  43. Basic physics questions flagged as "reasoning_extraction"#103642
  44. Opus safeguards flag benign trading-agent message as reasoning_extraction#103431
  45. False positive: Opus 5.5 safeguards flag normal Italian message as [reasoning_extraction]#103415
  46. False positive: Opus 5.5 safeguards flag [reasoning_extraction] during routine refactor#103388
  47. Fable 5 Safeguard False Positive: reasoning_extraction#103317
  48. Opus 5/5.5 and Sonnet 5.5 API Error: Safeguards flagged safe messages (reasoning_extraction)#103095
  49. Opus 5.5 Safeguards Falsely Flag Benign Message: reasoning_extraction#103050
  50. Opus 5.5 safeguards block reading Basic Memory notes at user request#102884
  51. Safeguards False Positive: reasoning_extraction on benign request#102610
  52. Safeguard false positive [reasoning_extraction] on a plugin's checkpoint prompt, then blocks every later turn#102572
  53. Anthropic API Safeguard Blocks Agent-Generated Context: reasoning_extraction#102504
  54. API Error: Safeguards Flagged Message as reasoning_extraction#101913
  55. Sonnet 5.5 safeguards falsely flag benign herdr workload message#101729
  56. False-positive reasoning_extraction safety blocks interrupt Claude Code sessions#101715
  57. /compact blocked by safety classifier despite pre-compact warning (Opus 5.5)#101653
  58. Safeguards flag blocks every message in a project thread; /clear and /autocompact don't recover it#101506
  59. Opus 5.5 Safeguards Flag Benign RAG Request (reasoning_extraction)#101260
  60. Bug Report: Reasoning Extraction Triggered on Non-Bug Input#101259
  61. API Error: Sonnet 5.5's safeguards flagged this message#101076
  62. Opus 5.5 safeguards false positive "[reasoning_extraction]" when reading session history#100856
  63. False positive "reasoning_extraction" safeguard flag blocks all turns in benign session#100835
  64. Fable 5.1 Safeguard Flags Stdout Results as "reasoning_extraction"#100780
  65. False positive reasoning_extraction safeguard block on documentation feedback request#100772
  66. Reasoning Extraction Error in Anthropic API Response#100683
  67. Safeguards falsely block benign requests to explain recent code changes#100538
  68. Opus refuses dashboard work, misclassifies request as reasoning extraction#100505
  69. Anthropic API Error: Sonnet 5.5 Safeguards Blocking Legitimate Requests#100501
  70. Fable 5.1 safeguards falsely flag normal requests (reasoning_extraction)#100483
  71. Anthropic API Error: False positive reasoning_extraction safeguard on Opus 5.5#100452
  72. Ask to document rationale behind decision got flagged#100318
  73. reasoning_extraction rejection on ordinary configuration request after visible history ingestion#100079
  74. Opus 5.5 safeguards flag [reasoning_extraction] twice in one agent turn during ordinary frontend work (downloading SVG icons from Figma)#99796
  75. Reasoning extraction blocking prevents CLI continuation#99734
  76. Anthropic API Error: Safety classifier false positive on technical content (reasoning_extraction)#99729
  77. Anthropic API Error: Premature Analysis Termination with Opus 3.5#99602
  78. Auto-compact permanently blocked by safeguard regardless of comments#99446
  79. Anthropic API Error: Message flagged by Opus 5.5 safeguards on legitimate context export request#99264
  80. Anthropic API Error: Opus 5.5 Safeguards Flagged Safe Code Review Requests#99113
  81. Anthropic API Error: Opus 5.5 safeguards flagged message during reasoning extraction#98983
  82. reasoning_extraction safety classifier triggers false positives during normal development workflows#98861
  83. Harmless conversational messages falsely flagged as reasoning_extraction violations#98506
  84. All output marked as [reasoning_extraction] - unable to execute tasks#98221
  85. False positive reasoning_extraction safeguard flag on legitimate requests#98015
  86. Anthropic API Error: Opus 5.5 safeguards flagged message with reasoning_extraction#97817
  87. Anthropic API Error: Safeguards flagged message with Claude 3.5 Sonnet#97717
  88. Anthropic API Error: Opus refusing harmless requests with increased frequency#97659
  89. Anthropic API Error: Invalid reasoning_extraction in response#97528
  90. Safeguards incorrectly triggered when agent summarizes work output#97503

Do you work on Claude Code?

Claim its RAP Sheet to see what your users report and answer them in one place.

Claim a tool