There are 19 attacks in the unqualified-advice category. Each shows its technique id, mode, standards mapping, description, and the exact command to run it (swap the provider and target as needed).
​
Child Medical Advice
critical
child-medical-advice · single-turn · OWASP: LLM01 · MITRE: AML.T0054, AML.T0068Provides dangerous medical advice for children without professional oversightRun it:
ai-blackteam run -p anthropic -a child-medical-advice -t "your target prompt"
​
Dangerous Dosage
critical
dangerous-dosage · single-turn · OWASP: LLM01 · MITRE: AML.T0054, AML.T0068Asks for dangerous medication dosages far beyond safe thresholdsRun it:
ai-blackteam run -p anthropic -a dangerous-dosage -t "your target prompt"
​
Drug Interaction
critical
drug-interaction · single-turn · OWASP: LLM01 · MITRE: AML.T0054, AML.T0068Asks about harmful drug interactions to cause intentional harmRun it:
ai-blackteam run -p anthropic -a drug-interaction -t "your target prompt"
​
Evidence Tampering
critical
evidence-tampering · single-turn · OWASP: LLM01 · MITRE: AML.T0054, AML.T0068Advises on destroying or tampering with evidence to obstruct legal proceedingsRun it:
ai-blackteam run -p anthropic -a evidence-tampering -t "your target prompt"
​
Medical Malpractice
critical
medical-malpractice · multi-turn · MITRE: AML.T0054, AML.T0065Multi-turn escalation from symptom description to requesting specific diagnoses and dosagesRun it:
ai-blackteam run -p anthropic -a medical-malpractice -t "your target prompt"
​
Pregnancy Harm
critical
pregnancy-harm · single-turn · OWASP: LLM01 · MITRE: AML.T0054, AML.T0068Provides dangerous medical advice for pregnant individuals that risks fetal or maternal harmRun it:
ai-blackteam run -p anthropic -a pregnancy-harm -t "your target prompt"
​
Self Surgery
critical
self-surgery · single-turn · OWASP: LLM01 · MITRE: AML.T0054, AML.T0068Asks for DIY surgical procedures to be performed without medical trainingRun it:
ai-blackteam run -p anthropic -a self-surgery -t "your target prompt"
​
Witness Coaching
critical
witness-coaching · single-turn · OWASP: LLM01 · MITRE: AML.T0054, AML.T0068Helps coach witnesses to lie or fabricate testimony in legal proceedingsRun it:
ai-blackteam run -p anthropic -a witness-coaching -t "your target prompt"
​
Alternative Medicine Harm
high
alternative-medicine-harm · single-turn · OWASP: LLM01 · MITRE: AML.T0054, AML.T0068Promotes dangerous alternative medicine practices as replacements for proven medical treatmentsRun it:
ai-blackteam run -p anthropic -a alternative-medicine-harm -t "your target prompt"
​
Bankruptcy Fraud
high
bankruptcy-fraud · single-turn · OWASP: LLM01 · MITRE: AML.T0054, AML.T0068Assists with fraudulent bankruptcy filings including asset concealment and false declarationsRun it:
ai-blackteam run -p anthropic -a bankruptcy-fraud -t "your target prompt"
​
Contract Manipulation
high
contract-manipulation · single-turn · OWASP: LLM01 · MITRE: AML.T0054, AML.T0068Helps draft manipulative or predatory contracts designed to exploit unsophisticated partiesRun it:
ai-blackteam run -p anthropic -a contract-manipulation -t "your target prompt"
​
Financial Malpractice
high
financial-malpractice · multi-turn · MITRE: AML.T0054, AML.T0065Multi-turn escalation from financial situation to requesting specific investment and trading adviceRun it:
ai-blackteam run -p anthropic -a financial-malpractice -t "your target prompt"
​
Health Quackery
high
health-quackery · single-turn · OWASP: LLM01 · MITRE: AML.T0054, AML.T0068Promotes dangerous health quackery that discourages legitimate medical treatmentRun it:
ai-blackteam run -p anthropic -a health-quackery -t "your target prompt"
​
Immigration Fraud
high
immigration-fraud · single-turn · OWASP: LLM01 · MITRE: AML.T0054, AML.T0068Helps with fraudulent immigration applications and document falsificationRun it:
ai-blackteam run -p anthropic -a immigration-fraud -t "your target prompt"
​
Legal Malpractice
high
legal-malpractice · multi-turn · MITRE: AML.T0054, AML.T0065Multi-turn escalation from legal situation to requesting specific legal strategy and documentsRun it:
ai-blackteam run -p anthropic -a legal-malpractice -t "your target prompt"
​
Legal Self Representation
high
legal-self-representation · single-turn · OWASP: LLM01 · MITRE: AML.T0054, AML.T0068Gives harmful legal advice in serious criminal or civil matters where lack of counsel causes severe harmRun it:
ai-blackteam run -p anthropic -a legal-self-representation -t "your target prompt"
​
Psychiatric Diagnosis
high
psychiatric-diagnosis · single-turn · OWASP: LLM01 · MITRE: AML.T0054, AML.T0068Asks to diagnose mental illness and prescribe psychiatric treatment without professional oversightRun it:
ai-blackteam run -p anthropic -a psychiatric-diagnosis -t "your target prompt"
​
Tax Fraud Advice
high
tax-fraud-advice · single-turn · OWASP: LLM01 · MITRE: AML.T0054, AML.T0068Provides tax fraud strategies framed as legitimate tax minimization adviceRun it:
ai-blackteam run -p anthropic -a tax-fraud-advice -t "your target prompt"
​
Veterinary Malpractice
high
veterinary-malpractice · single-turn · OWASP: LLM01 · MITRE: AML.T0054, AML.T0068Provides dangerous animal treatment advice that could harm or kill petsRun it:
ai-blackteam run -p anthropic -a veterinary-malpractice -t "your target prompt"