# 0DIN

> 0DIN (0Day Investigative Network) is Mozilla's GenAI bug bounty program and open AI security research platform. We coordinate disclosure of vulnerabilities in large language models, agentic systems, and other generative AI; publish open taxonomies and scoring frameworks for jailbreak research; and ship an open-source AI scanner and real-time threat feed for defenders.

All content under `0din.ai` is public and citable. Vulnerability disclosures, vendor and model statistics, scanner scorecards, and research methodology are intended for use by researchers, defenders, vendors, and journalists. Authenticated portals (admin, scanner, threat feed dashboards) are out of scope.

## Docs

- [Research](https://0din.ai/research.md): Index of 0DIN's public research on GenAI vulnerability classes, taxonomies, and evaluation methods.
- [QuickStart Guide](https://0din.ai/research/quick_start.md): How to submit a vulnerability to the bug bounty program.
- [Jailbreak Taxonomy](https://0din.ai/research/taxonomy.md): Hierarchical taxonomy of jailbreak categories, strategies, and techniques.
- [Jailbreak Taxonomy Reference Tool](https://0din.ai/research/taxonomy/reference.md): Browsable reference for the taxonomy with examples.
- [Jailbreak Evaluation Framework (JEF)](https://0din.ai/research/jef.md): Reproducible scoring framework for jailbreak severity.
- [Social Impact Score (SIS)](https://0din.ai/research/sis.md): Scoring system for societal harm from a given vulnerability.
- [Nude Imagery Rating System (NIRS)](https://0din.ai/research/nirs.md): Rating system for non-consensual or harmful imagery generation.

## Data

- [Disclosures](https://0din.ai/disclosures): Index of publicly disclosed vulnerabilities.
- [Example Disclosure: Multiple Model Guardrail Jailbreak via "Plagiarism Detection" Tactic](https://0din.ai/disclosures/0591c356-a3eb-4693-86e8-a5bda3d77876.md): Representative disclosure with metadata, taxonomy mappings, and detection signatures.
- [Leaderboard](https://0din.ai/leaderboard): Top contributing researchers ranked by accepted submissions.
- [Example Researcher Profile](https://0din.ai/researchers/7ef91915-53d2-4d1f-b15b-cfd30385f42f.md): Representative public researcher profile.

## Products

- [0DIN Scanner](https://0din.ai/marketing/scanner.md): Hosted AI security scanner for monitoring vulnerabilities in GenAI and agentic models.
- [0DIN Open-Source Scanner](https://0din.ai/marketing/open_source_scanner.md): Free, self-hosted scanner with 179 probes across 35 vulnerability families.
- [0DIN AI Vulnerability Intelligence](https://0din.ai/marketing/threat_intel.md): Real-time intelligence on emerging GenAI vulnerabilities and threats.
- [0DIN SusFactor Model](https://0din.ai/marketing/susfactor-model.md): Scoring of suspicious prompts using a local model, with trial access on request.
- [Bug Bounty](https://0din.ai/marketing/bug_bounty.md): Information for organizations sponsoring or participating in the program.
- [Program Scope](https://0din.ai/scope.md): Current models in scope for bug bounty submissions.
- [Program Scope: Apps](https://0din.ai/scope/apps.md): Current applications in scope for bug bounty submissions.
- [Datasheets](https://0din.ai/products/datasheets): Product datasheets (PDF) for download.

## Solutions

- [Attack Surface - Chatbots](https://0din.ai/solutions/chatbots.md): Stress-testing public and internal chat interfaces against researcher-validated exploits, mapped to OWASP LLM Top 10 and MITRE ATLAS.
- [Attack Surface - AI Applications](https://0din.ai/solutions/ai-apps.md): Testing copilots, RAG, and in-product LLM features for indirect injection, cross-context leakage, and downstream exploitation.
- [Attack Surface - Agents](https://0din.ai/solutions/agentic.md): Probing tool-using and autonomous agents for goal hijack, tool misuse, and unbounded behavior before production.

## Program

- [About 0DIN](https://0din.ai/about.md): Mission, team, and program overview.
- [Bug Bounty Program Terms and Conditions](https://0din.ai/policy.md): Eligibility, scope, payouts, and disclosure rules for researchers.
- [Privacy Notice](https://0din.ai/privacy.md): How 0DIN handles researcher and visitor data.

## Blog

- [Blog](https://0din.ai/blog): Index of research write-ups, vulnerability analyses, and program announcements.
- [Example Blog Post: SusFactor: Detecting Jailbreak and Prompt Injection Attacks Using Real-World Threat Intelligence](https://0din.ai/blog/susfactor-paper.md): Representative blog post in markdown.

## Disclosures

- [Multiple Model Guardrail Jailbreak via "Plagiarism Detection" Tactic](https://0din.ai/disclosures/0591c356-a3eb-4693-86e8-a5bda3d77876.md)
- [Multiple Model Guardrail Jailbreak via "LISP Documentation Transform" Tactic](https://0din.ai/disclosures/96f7c09e-c182-42ea-8d7e-8484c18306ce.md)
- [Multiple Model Guardrail Jailbreak via "Inverted Rules Chemistry Video" Tactic](https://0din.ai/disclosures/184f5586-5425-4a61-aced-1b1fac275967.md)
- [Multiple Model Guardrail Jailbreak via "Educational Research Document" Tactic](https://0din.ai/disclosures/9b51c324-60f1-4807-ad50-9d59ad37d444.md)
- [Multiple Model Guardrail Jailbreak via "Dialect Transition" Tactic](https://0din.ai/disclosures/aabdc2e9-46e4-40db-a6a8-1c092bdc24db.md)
- [Multiple Model Guardrail Jailbreak via "Interactive Lab Worksheet" Tactic](https://0din.ai/disclosures/2795ad5a-0c15-45cf-817d-69384cdeea89.md)
- [Multiple Model Guardrail Jailbreak via "Entropy Compression Schema" Tactic](https://0din.ai/disclosures/67eb68a6-8bb1-4206-b301-44ebe65816bb.md)
- [Multiple Model Guardrail Jailbreak via "Character Substitution" Tactic](https://0din.ai/disclosures/fc8cec5d-79e2-4952-a727-64e7c6a579c6.md)
- [Multiple Model Guardrail Jailbreak via "Anglo-Saxon Constraint" Tactic](https://0din.ai/disclosures/495fb676-9470-43ee-b707-b10d9ea02495.md)
- [X.ai Grok 2/3 Guardrail Jailbreak via "Chaotic Formatting" Tactic](https://0din.ai/disclosures/878a8b17-8a77-4493-9b2a-9856254785af.md)
- [Multiple Model Guardrail Jailbreak via Amino Acid Encoding Tactic](https://0din.ai/disclosures/bb9e653c-0041-42e7-b79c-a82f9d6c1e1f.md)
- [Multiple Model Guardrail Jailbreak via "Verbose Chemical Simulation" Tactic](https://0din.ai/disclosures/4266dbd4-50e2-4dac-ad2f-f1042895e0ee.md)
- [Multiple Model Guardrail Jailbreak via "Notion Template" Tactic](https://0din.ai/disclosures/424c8e32-f265-4957-8ebe-df2ec3ddf888.md)
- [Multiple Model Guardrail Jailbreak via "Academic Lab Protocol Manuscript" Tactic](https://0din.ai/disclosures/3c9f5962-23c1-4066-83de-8f0269e2fae7.md)
- [Multiple Model Guardrail Jailbreak via "Hiring Test Marking Scheme" Tactic](https://0din.ai/disclosures/395790b3-e2b4-418c-9938-b5f99e0d5ef2.md)
- [Multiple Model Guardrail Jailbreak via "Chemical Inquiry Escalation" Tactic](https://0din.ai/disclosures/1cf46993-189b-45c5-a1d3-a3f3fee8df31.md)
- [Multiple Model Guardrail Jailbreak via "Weighted Reaction Graph" Tactic](https://0din.ai/disclosures/49bf0056-a0d0-49f9-941f-86bbe2c6a7c9.md)
- [Multiple Model Guardrail Jailbreak via "Chemistry Challenge Framework" Tactic](https://0din.ai/disclosures/4f917ffa-b353-4d07-8fb4-67dc13218db8.md)
- [Multiple Model Guardrail Jailbreak via "Forensic Training Scenario" Tactic](https://0din.ai/disclosures/86117277-6c3e-4109-8721-5fe5b2e49b22.md)
- [Multiple Model Guardrail Jailbreak via "Escalating Academic Deep Dive" Tactic](https://0din.ai/disclosures/88ae3b65-8147-444b-a43f-7bc34da3bc77.md)
- [Multiple Model Guardrail Jailbreak via "Patent Litigation Framing" Tactic](https://0din.ai/disclosures/8e6c425b-1b60-4577-b1f8-4cacc4cb8d5e.md)
- [Mistral LeChat Data Exfiltration via Email Embedded Indirect Prompt Injection](https://0din.ai/disclosures/f80b23c1-26cb-4484-8ea2-0e7a82867931.md)
- [Multiple Model Guardrail Jailbreak via "Analytical Log Forensics" Tactic](https://0din.ai/disclosures/f809714d-4998-42eb-ad5f-1f5a930ed0a4.md)
- [Multiple Model Guardrail Jailbreak via "Educational Lab Inventory Simulator" Tactic](https://0din.ai/disclosures/61e40ed6-a8c4-4566-a1b4-895cef8d38d6.md)
- [Multiple Model Guardrail Jailbreak via "Forensic Court Blueprint Persona" Tactic](https://0din.ai/disclosures/06982d0d-4ad1-4b0b-8a50-b4ea7853c7ca.md)
- [Multiple Model Guardrail Jailbreak via "Forensic Field Guide" Tactic](https://0din.ai/threatfeed/a50c962f-1a46-4382-b810-9a2b4c5c6091/public.md)
- [Multiple Model Guardrail Jailbreak via "Hex Recipe Book" Tactic](https://0din.ai/threatfeed/06b58763-8a7b-4d82-afb6-ebe738b378a4/public.md)
- [Multiple Model Prompt Injection via "Correction" Tactic](https://0din.ai/threatfeed/439777fa-377a-43cb-83c6-c1a012e384b1/public.md)
- [Multiple Model Guardrail Jailbreak via "End-of-Response Override" Tactic](https://0din.ai/threatfeed/4e8ed295-a772-4a29-8727-cbefada44123/public.md)
- [Multiple Model Guardrail Jailbreak via "Incremental Table Completion" Tactic](https://0din.ai/threatfeed/0d1e2015-f6ce-4b79-8bde-edd26bfc6f02/public.md)
- [Multiple Model Guardrail Jailbreak via "Placeholder Injection" Tactic](https://0din.ai/threatfeed/8ccd7f81-4e36-4084-8061-cec6e2d83ece/public.md)

## Optional

- [Jailbreak Evaluation Calculator](https://0din.ai/research/jef/calculator): Interactive calculator for JEF scoring.
- [Legal Archive](https://0din.ai/legal_archive.md): Archived versions of policies and terms.
- [Sitemap](/sitemaps/sitemap.xml.gz): Full URL index for crawling.