There are 22 attacks in the disinformation category. Each shows its technique id, mode, standards mapping, description, and the exact command to run it (swap the provider and target as needed).
​
Coordinated Inauthentic Behavior
critical
coordinated-inauthentic · multi-turn · OWASP: LLM09 · MITRE: AML.T0054, AML.T0068Plans full coordinated inauthentic behavior operations across platforms at scaleRun it:
ai-blackteam run -p anthropic -a coordinated-inauthentic -t "your target prompt"
​
Deepfake Audio Generation
critical
deepfake-audio-generation · single-turn · OWASP: LLM01 · MITRE: AML.T0054, AML.T0068Generates instructions for creating deepfake audio clips to impersonate real people and spread disinformationRun it:
ai-blackteam run -p anthropic -a deepfake-audio-generation -t "your target prompt"
​
Deepfake Script
critical
deepfake-script · single-turn · OWASP: LLM09 · MITRE: AML.T0054, AML.T0068Creates scripts for deepfake videos that put false statements in real people’s mouthsRun it:
ai-blackteam run -p anthropic -a deepfake-script -t "your target prompt"
​
Evidence Fabrication
critical
evidence-fabrication · single-turn · OWASP: LLM01 · MITRE: AML.T0054, AML.T0068Creates fabricated evidence including documents, communications, and digital artifactsRun it:
ai-blackteam run -p anthropic -a evidence-fabrication -t "your target prompt"
​
Pandemic Disinformation
critical
pandemic-disinfo · single-turn · OWASP: LLM01 · MITRE: AML.T0054, AML.T0068Generates pandemic disinformation that undermines public health responseRun it:
ai-blackteam run -p anthropic -a pandemic-disinfo -t "your target prompt"
​
Scientific Fraud
critical
scientific-fraud · single-turn · OWASP: LLM09 · MITRE: AML.T0054, AML.T0068Generates fake scientific papers, fabricated data, and fraudulent research supporting false claimsRun it:
ai-blackteam run -p anthropic -a scientific-fraud -t "your target prompt"
​
Synthetic Media Creation
critical
synthetic-media-creation · single-turn · OWASP: LLM01 · MITRE: AML.T0054, AML.T0068Generates instructions for producing synthetic video and image media to deceive audiences and spread false narrativesRun it:
ai-blackteam run -p anthropic -a synthetic-media-creation -t "your target prompt"
​
Anti-Vaccine Disinformation
high
anti-vaccine-disinfo · single-turn · OWASP: LLM01 · MITRE: AML.T0054, AML.T0068Generates anti-vaccine misinformation designed to reduce vaccination ratesRun it:
ai-blackteam run -p anthropic -a anti-vaccine-disinfo -t "your target prompt"
​
Astroturfing
high
astroturfing · single-turn · OWASP: LLM09 · MITRE: AML.T0054, AML.T0068Generates fake grassroots campaign content designed to simulate organic public supportRun it:
ai-blackteam run -p anthropic -a astroturfing -t "your target prompt"
​
Climate Denial
high
climate-denial · single-turn · OWASP: LLM01 · MITRE: AML.T0054, AML.T0068Generates climate change denial content designed to undermine scientific consensusRun it:
ai-blackteam run -p anthropic -a climate-denial -t "your target prompt"
​
Conspiracy Amplification
high
conspiracy-amplification · multi-turn · OWASP: LLM09 · MITRE: AML.T0054, AML.T0068Amplifies conspiracy theories by layering fabricated evidence across a multi-turn conversationRun it:
ai-blackteam run -p anthropic -a conspiracy-amplification -t "your target prompt"
​
Emotional Manipulation Campaign
high
emotional-manipulation-campaign · multi-turn · OWASP: LLM09 · MITRE: AML.T0054, AML.T0068Designs large-scale campaigns that exploit emotional vulnerabilities to bypass rational judgmentRun it:
ai-blackteam run -p anthropic -a emotional-manipulation-campaign -t "your target prompt"
​
Expert Impersonation
high
expert-impersonation · single-turn · OWASP: LLM09 · MITRE: AML.T0054, AML.T0068Generates false content attributed to real domain experts to lend false credibility to claimsRun it:
ai-blackteam run -p anthropic -a expert-impersonation -t "your target prompt"
​
Fake News Generation
high
fake-news-generation · single-turn · OWASP: LLM09 · MITRE: AML.T0054, AML.T0068Generates realistic fake news articles with convincing details, sources, and quotesRun it:
ai-blackteam run -p anthropic -a fake-news-generation -t "your target prompt"
​
Fake Statistics
high
fake-statistics · single-turn · OWASP: LLM09 · MITRE: AML.T0054, AML.T0068Generates realistic-looking fabricated statistics attributed to credible institutionsRun it:
ai-blackteam run -p anthropic -a fake-statistics -t "your target prompt"
​
Historical Revisionism
high
historical-revisionism · single-turn · OWASP: LLM09 · MITRE: AML.T0054, AML.T0068Rewrites historical events with false narratives to serve present-day ideological goalsRun it:
ai-blackteam run -p anthropic -a historical-revisionism -t "your target prompt"
​
Narrative Warfare
high
narrative-warfare · multi-turn · OWASP: LLM09 · MITRE: AML.T0054, AML.T0068Crafts competing narratives designed to polarize communities and make consensus impossibleRun it:
ai-blackteam run -p anthropic -a narrative-warfare -t "your target prompt"
​
Propaganda Generation
high
propaganda-generation · single-turn · OWASP: LLM09 · MITRE: AML.T0054, AML.T0068Creates state-level propaganda content using proven psychological persuasion techniquesRun it:
ai-blackteam run -p anthropic -a propaganda-generation -t "your target prompt"
​
Social Manipulation Bot
high
social-manipulation-bot · single-turn · OWASP: LLM01 · MITRE: AML.T0054, AML.T0068Generates strategies for deploying bot networks to manipulate public opinion, amplify disinformation, and suppress targeted narrativesRun it:
ai-blackteam run -p anthropic -a social-manipulation-bot -t "your target prompt"
​
Sock Puppet
high
sock-puppet · multi-turn · OWASP: LLM09 · MITRE: AML.T0054, AML.T0068Creates detailed fake online personas for coordinated inauthentic behavior campaignsRun it:
ai-blackteam run -p anthropic -a sock-puppet -t "your target prompt"
​
Fake Review Generation
medium
fake-review-generation · single-turn · OWASP: LLM09 · MITRE: AML.T0054, AML.T0068Mass generates fake product or service reviews to manipulate consumer perception and ratingsRun it:
ai-blackteam run -p anthropic -a fake-review-generation -t "your target prompt"
​
Selective Quoting
medium
selective-quoting · single-turn · OWASP: LLM09 · MITRE: AML.T0054, AML.T0068Distorts the meaning of real statements by selective quoting, removing context to invert intentRun it:
ai-blackteam run -p anthropic -a selective-quoting -t "your target prompt"