#ProtectAI
#Identiverse 2025 Keynotes continue this morning with
Authenticating the Future: AI's Role in Identity's Next Chapter - Diana Kelley, CISO, ProtectAI
from Diana Kelley - CISO of Protect AI
June 5, 2025 at 4:19 PM
March 12, 2026 at 5:28 AM
August 16, 2025 at 11:19 PM
Vulnhuntr leverages the power of LLMs to automatically create and analyze entire code call chains starting from remote user input and ending at server output for detection of complex, multi-step, security-bypassing vulnerabilities github.com/protectai/vu...
GitHub - protectai/vulnhuntr: Zero shot vulnerability discovery using LLMs
Zero shot vulnerability discovery using LLMs. Contribute to protectai/vulnhuntr development by creating an account on GitHub.
github.com
December 19, 2024 at 10:24 PM
AI prompt injection detection comparison (March 2026):

No protection: 0/91 blocked
Keyword filter: 27/91
ProtectAI DeBERTa: 62/91
SCBE-AETHERMOORE: 91/91, 0 false positives

Open source. Patent pending.

pip install scbe-aethermoore
https://aethermoorgames.com/redteam.html
March 25, 2026 at 1:38 PM
Red team test suites now public on HuggingFace.

91 attacks. 10 categories. 0 false positives.
Beat ProtectAI DeBERTa v2 (91/91 vs 62/91).

Null space attack fingerprinting. Session tracking. Semantic tongue detection.

https://huggingface.co/datasets/issdandavis/scbe-red-team-benchmarks

Patent pen
March 25, 2026 at 8:30 AM
We built an AI governance framework that uses hyperbolic geometry instead of text rules.

91/91 attacks blocked. 0 false positives. Beats ProtectAI DeBERTa v2.

NIST AI RMF: 23/23 compliance checks pass.

Open source + patent pending.

https://github.com/issdandavis/SCBE-AETHERMOORE
March 25, 2026 at 4:38 AM
A collection of real world AI/ML exploits for responsibly disclosed vulnerabilities.
GitHub - protectai/ai-exploits: A collection of real world AI/ML exploits for responsibly disclosed ...
A collection of real world AI/ML exploits for responsibly disclosed vulnerabilities - GitHub - protectai/ai-exploits: A collection of real world AI/ML exploits for responsibly disclosed vulnerabil...
github.com
November 16, 2023 at 5:57 PM
Parts 1-4/5 — EFTA01307434.jpg
#epsteinweb #efta01307434
https://epsteinweb.org
February 17, 2026 at 4:42 AM
Parts 1-4/5 — EFTA01307238.jpg
#epsteinweb #efta01307238
https://epsteinweb.org
February 17, 2026 at 4:31 AM
Parts 1-4/5 — EFTA01306972.jpg
#epsteinweb #efta01306972
https://epsteinweb.org
February 17, 2026 at 4:16 AM
Parts 1-4/5 — EFTA01306908.jpg
#epsteinweb #efta01306908
https://epsteinweb.org
February 17, 2026 at 4:12 AM
<a href="https://github.com/protectai/vulnhuntr" class="hover:underline text-blue-600 dark:text-sky-400 no-card-link" target="_blank" rel="noopener" data-link="bsky">github.com/protectai/vulnhun...
GitHub - protectai/vulnhuntr: Zero shot vulnerability discovery using LLMs
GitHub - protectai/vulnhuntr: Zero shot vulnerability discovery using LLMs
github.com
May 4, 2025 at 3:13 AM
Furthermore, when evaluated on public benchmarks, it consistently outperforms strong baselines like protectai/deberta-v3-base-prompt-injection-v2. This work details Sentinel's architecture, its meticulous dataset curation, its training methodology, [5/6 of https://arxiv.org/abs/2506.05446v1]
June 9, 2025 at 5:58 AM
malicious prompts, and evaluate these tools performance against a baseline LLM model (ChatGPT-3.5-Turbo). Our results show that the baseline model has too many false positives to be used for this task. Lakera Guard and ProtectAI LLM Guard emerged as [5/7 of https://arxiv.org/abs/2505.13028v1]
May 20, 2025 at 6:05 AM
May 2, 2025 at 4:59 PM