| --- |
| title: README |
| emoji: 🛡️ |
| colorFrom: gray |
| colorTo: red |
| sdk: static |
| pinned: false |
| --- |
| |
| <div style="text-align: center; margin-bottom: 1.5em;"> |
| <h1>🛡️ Astroware Inc.</h1> |
| <p><strong>AI Security Research & Agentic Safety</strong></p> |
| <p> |
| <a href="https://astroware.ai" style="text-decoration: none;">🌐 Website</a> | |
| <a href="https://github.com/astroware" style="text-decoration: none;">💻 GitHub</a> | |
| <a href="https://twitter.com/astroware" style="text-decoration: none;">🐦 Twitter</a> |
| </p> |
| </div> |
| |
| --- |
|
|
| ## Who We Are |
|
|
| Astroware is an **AI security startup** focused on safety, alignment, and agentic security research. We build tools and models that make AI systems safer to deploy at scale, with a particular focus on **guard models** and **constitutional AI classifiers** that act as runtime security layers for AI agents. |
|
|
| We are a Delaware C-Corp with a globally distributed team spanning Dubai and Bengaluru. |
|
|
| ## What We're Working On |
|
|
| ### 🔒 Guard Models & Constitutional Classifiers |
|
|
| Our core research area. We develop guard models that serve as runtime security layers for AI agents, preventing jailbreaks, prompt injection, and unsafe behavior. Our classifiers are built on a constitutional AI framework with structured severity tiers across harmful and benign behavioral categories. |
|
|
| ### 🔱 Trishool — Agentic Security Network |
|
|
| Trishool is our adversarial evaluation and agentic security platform, live on [Bittensor Subnet 23](https://bittensor.com). It stress-tests AI agents and guard models through real-world adversarial challenges, including jailbreak attacks against protected AI systems. Trishool's Phase 2 focuses on positioning guard models as the critical runtime defense layer for autonomous AI agents. |
|
|
| ### ⚖️ Alignment Research |
|
|
| We conduct alignment training research for large language models, including constitutional frameworks, severity-tiered taxonomies, and structured datasets for supervised fine-tuning and reinforcement learning. |
|
|
| ## Our Focus Areas |
|
|
| - **Guard model development** for runtime AI agent protection |
| - **Adversarial red-teaming** and jailbreak evaluation |
| - **Constitutional AI** classifiers and alignment frameworks |
| - **Agentic security** for multi-agent and autonomous systems |
| - **Open-source contributions** to AI safety tooling |
|
|
| ## Open Source |
|
|
| We believe AI security benefits from open collaboration. We actively contribute to open-source AI safety projects and publish our guard model research, adversarial evaluation tools, and security architectures for the community to build on. |
|
|
| --- |
|
|
| <p style="text-align: center; color: #888; font-size: 0.9em;"> |
| Building the security layer for the agentic AI era. 🚀 |
| </p> |
