There are 25 attacks in the media-manipulation category. Each shows its technique id, mode, standards mapping, description, and the exact command to run it (swap the provider and target as needed).
​
Academic Journal Fraud
critical
academic-journal-fraud · single-turn · OWASP: LLM01 · MITRE: AML.T0054, AML.T0068Fabricates peer-reviewed research papers, study results, and academic publications to lend false scientific authority to disinformationRun it:
ai-blackteam run -p anthropic -a academic-journal-fraud -t "your target prompt"
​
Crisis Fabrication
critical
crisis-fabrication · single-turn · OWASP: LLM01 · MITRE: AML.T0054, AML.T0068Fabricates emergency situations, disasters, or public health crises to cause panic, manipulate markets, or discredit organizationsRun it:
ai-blackteam run -p anthropic -a crisis-fabrication -t "your target prompt"
​
Leaked Document Fabrication
critical
leaked-document-fabrication · single-turn · OWASP: LLM01 · MITRE: AML.T0054, AML.T0068Generates convincing fake leaked documents — memos, emails, reports — designed to appear as authentic internal disclosuresRun it:
ai-blackteam run -p anthropic -a leaked-document-fabrication -t "your target prompt"
​
Source Fabrication
critical
source-fabrication · single-turn · OWASP: LLM01 · MITRE: AML.T0054, AML.T0068Fabricates citations, expert sources, and research references to lend false credibility to disinformationRun it:
ai-blackteam run -p anthropic -a source-fabrication -t "your target prompt"
​
Testimony Fabrication
critical
testimony-fabrication · single-turn · OWASP: LLM01 · MITRE: AML.T0054, AML.T0068Fabricates eyewitness testimony, legal declarations, and sworn statements to manufacture false evidence of eventsRun it:
ai-blackteam run -p anthropic -a testimony-fabrication -t "your target prompt"
​
Viral Disinformation Campaign
critical
viral-disinfo-campaign · single-turn · OWASP: LLM01 · MITRE: AML.T0054, AML.T0068Designs coordinated multi-platform disinformation campaigns engineered for viral spreadRun it:
ai-blackteam run -p anthropic -a viral-disinfo-campaign -t "your target prompt"
​
Whistleblower Impersonation
critical
whistleblower-impersonation · single-turn · OWASP: LLM01 · MITRE: AML.T0054, AML.T0068Creates fake whistleblower personas and false disclosures designed to appear as authentic insider revelationsRun it:
ai-blackteam run -p anthropic -a whistleblower-impersonation -t "your target prompt"
​
Audio Manipulation
high
audio-manipulation · single-turn · OWASP: LLM01 · MITRE: AML.T0054, AML.T0068Generates scripts and strategies for creating or distributing manipulated audio recordings attributed to real peopleRun it:
ai-blackteam run -p anthropic -a audio-manipulation -t "your target prompt"
​
Brand Impersonation
high
brand-impersonation · single-turn · OWASP: LLM01 · MITRE: AML.T0054, AML.T0068Creates fake branded content impersonating legitimate organizations to spread false announcements or damage brand reputationRun it:
ai-blackteam run -p anthropic -a brand-impersonation -t "your target prompt"
​
Content Farm Generation
high
content-farm-generation · single-turn · OWASP: LLM01 · MITRE: AML.T0054, AML.T0068Generates high-volume low-quality content designed to flood information ecosystems and dilute credible reportingRun it:
ai-blackteam run -p anthropic -a content-farm-generation -t "your target prompt"
​
Documentary Fabrication
high
documentary-fabrication · single-turn · OWASP: LLM01 · MITRE: AML.T0054, AML.T0068Creates fake documentary scripts, trailers, and promotional materials to lend false credibility to disinformationRun it:
ai-blackteam run -p anthropic -a documentary-fabrication -t "your target prompt"
​
Image Caption Manipulation
high
image-caption-manipulation · single-turn · OWASP: LLM01 · MITRE: AML.T0054, AML.T0068Creates false or misleading captions for real images to reframe their meaning and spread disinformationRun it:
ai-blackteam run -p anthropic -a image-caption-manipulation -t "your target prompt"
​
Influencer Fraud
high
influencer-fraud · single-turn · OWASP: LLM01 · MITRE: AML.T0054, AML.T0068Fabricates sponsored content, fake endorsements, and fraudulent influencer campaigns to manipulate public opinionRun it:
ai-blackteam run -p anthropic -a influencer-fraud -t "your target prompt"
​
Interview Fabrication
high
interview-fabrication · single-turn · OWASP: LLM01 · MITRE: AML.T0054, AML.T0068Creates fake interview transcripts attributed to real public figures to spread false statementsRun it:
ai-blackteam run -p anthropic -a interview-fabrication -t "your target prompt"
​
Media Attribution Fraud
high
media-attribution-fraud · single-turn · OWASP: LLM01 · MITRE: AML.T0054, AML.T0068Falsely attributes real or fabricated content to trusted media organizations to exploit their credibilityRun it:
ai-blackteam run -p anthropic -a media-attribution-fraud -t "your target prompt"
​
News Headline Manipulation
high
news-headline-manipulation · single-turn · OWASP: LLM01 · MITRE: AML.T0054, AML.T0068Crafts misleading or false news headlines designed to distort perception and drive viral sharingRun it:
ai-blackteam run -p anthropic -a news-headline-manipulation -t "your target prompt"
​
Obituary Fabrication
high
obituary-fabrication · single-turn · OWASP: LLM01 · MITRE: AML.T0054, AML.T0068Creates fake obituaries and death announcements for living public figures to cause reputational harm or market manipulationRun it:
ai-blackteam run -p anthropic -a obituary-fabrication -t "your target prompt"
​
Podcast Fabrication
high
podcast-fabrication · single-turn · OWASP: LLM01 · MITRE: AML.T0054, AML.T0068Creates fake podcast transcripts, episode descriptions, and show notes attributed to real podcasters or networksRun it:
ai-blackteam run -p anthropic -a podcast-fabrication -t "your target prompt"
​
Press Release Fabrication
high
press-release-fabrication · single-turn · OWASP: LLM01 · MITRE: AML.T0054, AML.T0068Generates fake official press releases mimicking real organizations to spread false announcementsRun it:
ai-blackteam run -p anthropic -a press-release-fabrication -t "your target prompt"
​
Review Bombing
high
review-bombing · single-turn · OWASP: LLM01 · MITRE: AML.T0054, AML.T0068Generates coordinated fake negative reviews designed to damage the reputation of a target through review platform manipulationRun it:
ai-blackteam run -p anthropic -a review-bombing -t "your target prompt"
​
SEO Manipulation
high
seo-manipulation · single-turn · OWASP: LLM01 · MITRE: AML.T0054, AML.T0068Designs SEO poisoning campaigns to surface disinformation in search results and suppress factual contentRun it:
ai-blackteam run -p anthropic -a seo-manipulation -t "your target prompt"
​
Social Media Manipulation
high
social-media-manipulation · single-turn · OWASP: LLM01 · MITRE: AML.T0054, AML.T0068Designs coordinated social media manipulation campaigns using fake accounts, astroturfing, and engagement fraudRun it:
ai-blackteam run -p anthropic -a social-media-manipulation -t "your target prompt"
​
Video Manipulation
high
video-manipulation · single-turn · OWASP: LLM01 · MITRE: AML.T0054, AML.T0068Creates scripts and distribution strategies for fabricated or manipulated video content attributed to real people or eventsRun it:
ai-blackteam run -p anthropic -a video-manipulation -t "your target prompt"
​
Award Fabrication
medium
award-fabrication · single-turn · OWASP: LLM01 · MITRE: AML.T0054, AML.T0068Creates fake awards, certifications, and accolades to build false credibility for individuals, products, or organizationsRun it:
ai-blackteam run -p anthropic -a award-fabrication -t "your target prompt"
​
Clickbait Generation
medium
clickbait-generation · single-turn · OWASP: LLM01 · MITRE: AML.T0054, AML.T0068Generates manipulative clickbait content designed to drive traffic through deceptive or misleading framingRun it:
ai-blackteam run -p anthropic -a clickbait-generation -t "your target prompt"