Istiaque Ahmed

23/09/2026

After overdrinking, you do not care about your running speed or competing with anyone; you just run into the distance where no one wishes to explore, without even caring about Professorship.

23/09/2026

On weekdays, I usually start working at 9 AM, but on weekends, I start at 6 AM.

23/09/2026

18/09/2026

Better to die fighting than live in defeat.

16/09/2026

Introducing AgentRed: An Open-Source Red-Teaming Framework for Security Assessment of AI Agents

I am excited to announce the initial release of AgentRed, an open-source framework designed to help researchers and developers evaluate the security of AI agents.

As AI agents become more autonomous and gain the ability to use tools, access data, and interact with external systems, ensuring their security and reliability becomes increasingly important. AgentRed aims to provide a foundation for systematic security assessment and red-teaming of agentic AI systems.

🔍 AgentRed focuses on:

🛡️ Security testing of AI agents
🔴 Red-team inspired vulnerability assessment
🤖 Evaluation of agent behaviors and potential risks
🔧 Extensible framework design for future attack modules and security analysis

The project is now available as:

⭐ GitHub Repository
https://github.com/istiaque010/AgentRed

📦 Python Package (PyPI)
https://pypi.org/project/agentred/

Installation:

pip install agentred

This is the beginning of an ongoing effort. I welcome researchers, developers, and AI security enthusiasts to explore the framework, provide feedback, and contribute to improving the security of next-generation AI agent systems.

15/09/2026

I am pleased to share our latest work:

📄 "Transparent Identity Verification Approach Using MPC and Efficient Credential Status Handling"
🔗 https://arxiv.org/abs/2609.14195

In this paper, we propose a privacy-preserving and scalable identity verification framework using Multi-Party Computation (MPC), blockchain anchoring, and efficient credential status management.

Key contributions include:
🔹 Private off-chain computation with runtime proofs
🔹 Efficient credential status updates and compression
🔹 Reusable Verifiable Presentations with selective disclosure
🔹 Blockchain-based transparency and post-quantum security support

This work aims to advance secure, scalable, and future-proof digital identity infrastructures.

Many thanks to my co-authors:
👏 Shoji Kasahara
👏 Kentaroh Toyoda
👏 Tadashi Nakano
👏 Thi Hong Tran

A secure and privacy-preserving identity verification process is essential for digital ecosys- tems. Current eKYC frameworks that rely on Zero-Knowledge Proofs (ZKPs) face high computational cost, rigid circuit design, complex integration, and expensive on-chain verification. The W3C 2021 BitString-...

14/09/2026

🚀 Introducing AgentRed: An Open-Source Red-Teaming Framework for Security Assessment of AI Agents

I am excited to introduce AgentRed, an open-source framework I have developed to evaluate and improve the security of AI Agents.

As AI Agents become more autonomous and capable of interacting with external tools, ensuring their safety against threats such as prompt injection, tool misuse, and sensitive data leakage is becoming increasingly important.

AgentRed enables systematic security testing of AI Agents through adversarial evaluation and vulnerability assessment, helping developers identify security risks before real-world deployment.

🔐 My goal is to contribute toward building AI Agents that are more secure, trustworthy, and controllable.

The project will be released publicly soon. Stay tuned!

14/09/2026

It would not be a prerequisite for downloading the DB.

08/09/2026

“Yesterday, we built software for the hardware we had.
Today, we define the purpose, shape the features, and choose the hardware that fits.”

— A shift from hardware-driven development to purpose-driven technology.

29/08/2026

We struggle. We build. We grow — LessonQ