Its theme's rank, 14 days: 2 on the Claude Code sheet
47 people reported
All time, since Sep 15, including 3 thumbs-up on GitHub.
Safeguards block requests
“wrong safeguards flagged for biology research”
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/3603/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
Safeguards flag sessions and block responses to requests. Claude Code shows an API error that "safeguards flagged this session" with a Details tag such as [bio]; rewriting the prompt did not change anything.
Is this your bug?
File my report- Where
- API client
- Error text
API Error: Opus 5.5's safeguards flagged this session (https://www.anthropic.com/legal/aup).- Versions
- 2.1.272 to 2.1.314
- Platforms
- Linux, macOS, Windows
- How bad
- Breaks a normal workflow.
Part of: Safeguards block session requests34 bugs
- Cyber safeguard blocks sessions377 people
- Reasoning_extraction safeguard block134 people
- Cyber safeguard pauses sessions52 people
- Safeguards refuse work requests48 people
- Classifier halts routine replies12 people
- Requests blocked with a [cyber] flag8 people
- Refusal on own-code hardening11 people
- SSH to own devices refused7 people
- Pentest work blocked6 people
- CTF requests flagged as malicious6 people
Show all 33Show only the largest 10
- Classifier blocks own-project work6 people
- Classifier halts reverse engineering5 people
- Safeguards flag verified cyber work6 people
- Firmware requests flagged as cyber6 people
- Classifier stops benign replies5 people
- reasoning_extraction API error5 people
- Safeguard blocks all messages4 people
- [bio] classifier blocks responses4 people
- Reasoning extraction safeguard block4 people
- Security investigation requests blocked3 people
- Classifier stops DeFi bot work3 people
- Safeguards block CVP members3 people
- "[cyber]" flag stops sessions1 person
- Safety classifier blocks agent3 people
- Fictional hacking content flagged3 people
- Cyber verification prompt recurs2 people
- Cyber safeguards block own devices2 people
- Safety classifier stops auth work2 people
- Phishing research project flagged2 people
- Cyber stop on report tasks2 people
- Cyber safety block on audits2 people
- Biological safeguard triggers in sessions with no biology content2 people
- Unexpected safeguards notice2 people
Original reports46, newest first
- False positive: safety filter blocked a genomics ML brainstorm#107167
- API Error: Sonnet 5 Refused Request Citing AUP (bio)#107028
- False positive refusal on DRT2 bacterial defense system research question#106776
- Biosafety Classifier Too Sensitive, Flags Benign Requests#106283
- Biosafety classifier too sensitive, flags benign requests#106282
- Request Blocked for Local Web Vulnerability Scanning#106097
- Request incorrectly flagged for "Bio" content while building Discord bot API integration#106009
- Safeguards incorrectly flag legitimate biology research requests#105897
- Safety classifier halted a Write call summarizing a public drug-development blog post, with no reason given; /feedback also rejected#105585
- Bug Report: blocked by the bio guardrail#103627
Show all 46Show only the newest 10
- Opus 5.5 Safeguard False Positive [bio] on Science Education Video Task#103522
- Opus 5.5 Over-Triggers Safeguard on Graduate Molecular Toxicology Content#103430
- False-Positive [bio] Safeguard Silently Downgrades Opus 5.5 to Opus 5#103382
- Opus 5.5 [bio] safeguard false positives on routine life-sciences literature work, sharp increase since 2026-10-16#103049
- Refuses to test security vulnerability locally#102051
- Prompt classifier incorrectly blocks valid prompt#101795
- Opus 5.5 [bio] classifier interrupts benign biomedical literature work (reading published papers, pre-commit secret scan)#101274
- Opus 5.5 [bio] classifier flags benign in-silico protein-engineering session; blocks file-listing output and /feedback returns 403#101211
- [bio] safeguard blocks fantasy world-building session with no biology content#101086
- [bio] safeguard false positives repeatedly halt a protein-design session (76 flags, 40 request IDs); the stall is also misreported to the agent#100552
- Anthropic API: Excessive false positive safety flags on medical content#100352
- Anthropic API Error: Overly restrictive content filter on harmless bioinformatics operations#100287
- false-positive na "cybersecurity" promt to siema masz...#100015
- Unclear error flagging on present-tense capability questions#99963
- Local Project Application Violation Error#99601
- Inaccurate [bio] safety flag regarding legitimate university work making me unable to use Claude Code#99253
- Claude API Security Filter Triggering on Local SQL Injection Tests#99097
- Safeguards prevent testing of locally running applications#99061
- Overly restrictive content filtering for legitimate cancer research queries#98698
- Incorrect Content Policy Flagging: Open Source Library Wrapper Marked as "bio"#98602
- Anthropic API Error: Conversation Blocked by Biohazard Safeguards with Opus 5.5 and Sonnet 5.5#98346
- Classifier falsely flagging previously valid inputs#97700
- Safeguard triggers on local device testing#97351
- False Positive: Content Incorrectly Flagged#97277
- Anthropic API Error: Opus 5.5 safeguards incorrectly flagging legitimate biology research#96924
- classifier triggering on sending a '.' to wake claude after an api error.#96851
- False positive flagging without clear error context#96777
- Session terminated with incorrect
[bio]reason outside biology context#96730 - False Positive Prompt Flagging#96593
- Session paused with [bio] flag during legitimate metabolic modeling research#96457
- CVP Integration Fails Against Local Vulnerable Test Instances#96376
- Opus 5.5 refuses to summarize conversation, cites biology flag, suggests switching models#96275
- Model refuses to engage with biological/biotech content in CRUD delivery simulation tasks#95704
- False positive [bio] safeguard trigger on routine RNA-seq statistical analysis#95700
- Anthropic API Error: Model refuses request with vague "bio" classification#94693
- Anthropic API Error: Unexpected Refusal on Valid Request#94544