0DIN.ai

AI security programs mature when they move from opinions to operating loops. For 0DIN Scanner, the loop is clear: identify the target, run the test, capture the finding, decide whether it changes risk, and feed the result back into engineering or research. This is where tracking GenAI vulnerabilities as systems change matters. It gives CISOs, AppSec leaders, and AI platform owners a concrete reason to act instead of waiting for the next incident, audit, or board question. Use the post-launch window to create a baseline. Use every material model, prompt, tool, or workflow change as a reason to retest. Request a demo: https://lnkd.in/gaJUSX5C #0DIN #AISecurity #GenAISecurity #CISO

A mature AI security program needs a repeatable operating model. That model should help teams identify the target, test the boundary, record the result, prioritize the response, and retest when the system changes. 0DIN Open-Source Scanner fits into that workflow by focusing on free self-hosted AI security testing before deployment. The supporting capabilities are concrete: 179 probes, 35 vulnerability families, OWASP LLM Top 10 mapping, scheduled scans, PDF reports, and SIEM forwarding. This is the difference between knowing AI risk exists and having a process for reducing it. Try the scanner: https://lnkd.in/gS7RuQ7A #0DIN #OpenSource #AISecurity #LLMSecurity

0DIN Jailbreak Feed addresses a practical AI security problem: feeding SOC and risk teams actionable AI security context. For SOC teams, threat intel teams, AI risk owners, and CISOs, the value is not generic awareness. It is a more reliable way to create evidence, prioritize action, and keep up with changing AI systems. 0DIN Jailbreak Feed supports that by providing real-time alerts, behavior trends, vulnerability reports, probe context, source attribution, API access, and regulatory mapping. For security leaders, the important shift is cadence. AI systems change through model updates, prompt revisions, new tool access, and workflow expansion. Testing needs to follow that pace. Request a demo: https://lnkd.in/gaJUSX5C #0DIN #JailbreakFeed #ThreatIntel #AISecurity

AI security moves too fast for one-time reviews. 0DIN Jailbreak Feed gives teams a practical workflow for feeding SOC and risk teams actionable AI security context. The signal: real-time alerts, behavior trends, vulnerability reports, probe context, source attribution, API access, and regulatory mapping. Request a demo: https://lnkd.in/giucUEJ4 #0DIN #JailbreakFeed #ThreatIntel #AISecurity

AI security programs mature when they move from opinions to operating loops. For 0DIN Scanner, the loop is clear: identify the target, run the test, capture the finding, decide whether it changes risk, and feed the result back into engineering or research. This is where building a repeatable scan, remediate, retest loop matters. It gives CISOs, AppSec leaders, and AI platform owners a concrete reason to act instead of waiting for the next incident, audit, or board question. Use the post-launch window to create a baseline. Use every material model, prompt, tool, or workflow change as a reason to retest. Interested in seeing what 0DIN can do? Request a demo: https://lnkd.in/giucUEJ4 #0DIN #AISecurity #GenAISecurity #CISO

We just published a new 0DIN blog post on the Public Disclosures Corpus. The corpus is distributed weekly as a versioned JSONL dataset, making public GenAI vulnerability disclosures easier to use in research, evaluation, and defensive security workflows. It includes structured metadata across disclosures, with a vetted subset containing deeper technical intelligence such as triggering prompts, target responses, sector-specific variations, and detection signatures. This is about making AI security findings more accessible, reproducible, and useful for the people building defenses. Read it here: https://lnkd.in/g2rWuYKx

0DIN Scanner is for CISOs, AppSec leaders, and AI platform owners. Today's angle: turning AI governance into testable evidence. Why this matters: AI risk changes whenever a model updates, a prompt changes, an agent gets a new tool, or an app connects to sensitive data. Security teams need evidence that keeps up with that pace. 0DIN Scanner brings that work into a repeatable workflow through real-time vulnerability tracking, automated probe management, cross-model analysis, dashboards, and security reports. The operating model is simple: test the system, track the finding, prioritize remediation, and retest when models, prompts, tools, or workflows change. Request a demo: https://lnkd.in/giucUEJ4 #0DIN #AISecurity #GenAISecurity #CISO

We're proud to share that 0DIN.ai has partnered with the Mozilla Data Collective to open-source a dataset of our public disclosures - now live on Hugging Face. Our reasoning is simple: by open-sourcing a portion of our disclosures, we hope to empower the open-source community to approach LLM security with the same thoroughness and rigor we bring to our own customers. AI safety can't be locked behind enterprise walls. The community building the future of AI deserves the same tools, the same data, and the same starting line. https://lnkd.in/gDnBpCZv

The best part of the 0DIN bug bounty? Your finding doesn't just get fixed and forgotten. It becomes: - A probe in the Scanner (protecting everyone who scans) - An entry in the Threatfeed (informing security teams globally) - A public disclosure (building the industry's knowledge base) One submission. Multiplied impact. Join our Bug Bounty: https://lnkd.in/guzrhp-i #BugBounty #0DIN #Impact #SecurityResearch #AISecurity

Security Advisory - CVE-2026-41512 We've disclosed and patched a critical vulnerability (CVSS 9.9) in 0DIN AI Scanner affecting versions 1.0.0 through 1.4.0. A JavaScript injection flaw in our PlaywrightService allowed an authenticated user to execute arbitrary code in the scanner's Node.js runtime via a crafted URL breaking tenant isolation and exposing application secrets. Fixed in v1.4.1 - Please upgrade today. Huge thanks to Griffin Francis for the responsible disclosure and an outstanding technical write-up. This is what good security collaboration looks like and it's exactly why we open-sourced the scanner. Full advisory: https://lnkd.in/gwC97Xb6