| <!doctype html><html><head><meta charset=utf-8><title>Aligned Super Intelligence</title><style>body{font-family:Inter,system-ui,sans-serif;background:#0b0b0c;color:#e7e7e7;max-width:860px;margin:0 auto;padding:40px 24px;line-height:1.6} |
| h1{font-size:2.2em;margin:.2em 0;letter-spacing:-.02em}h2{border-top:1px solid #222;padding-top:1.2em;margin-top:1.6em;font-size:1.3em} |
| h3{color:#9aa0a6;font-weight:500}a{color:#e7e7e7}table{border-collapse:collapse;width:100%}td,th{border-bottom:1px solid #222;padding:8px;text-align:left} |
| code{background:#1a1a1c;padding:2px 6px;border-radius:4px}hr{border:0}sub{color:#777}div[align=center]{text-align:center}</style></head><body><div align="center"> |
| <h1>Aligned Super Intelligence</h1> |
| <h3>Offensive security for frontier AI systems</h3> |
| <p><em>We break AI systems before adversaries do, for the governments and institutions that can't afford to be second.</em></p> |
| <p><a href="https://alignedsafely.com"><strong>Website</strong></a> · <a href="https://huggingface.co/alignedsi/Aligned-1"><strong>Aligned-1</strong></a> · <a href="mailto:leadership@alignedsafely.com?subject=White-glove%20engagement"><strong>Engagements</strong></a></p> |
| </div> |
| <hr /> |
| <h2>Who we are</h2> |
| <p>Aligned Super Intelligence is an AI security lab. We are a team of hackers and researchers with industry security experience. We bring offensive security discipline to a new attack surface: large language models, retrieval pipelines, and autonomous agents that can read, write, and act.</p> |
| <p>As AI systems take on real authority, such as running tools, moving data, and making decisions, the cost of a single exploitable behavior goes up. Our job is to find those behaviors first, prove them with evidence, and help you close them.</p> |
| <h2>Who we work with</h2> |
| <table> |
| <thead> |
| <tr> |
| <th>Sector</th> |
| <th>Typical concern</th> |
| </tr> |
| </thead> |
| <tbody> |
| <tr> |
| <td><strong>Government & public sector</strong></td> |
| <td>AI deployed in citizen-facing and internal decision systems</td> |
| </tr> |
| <tr> |
| <td><strong>Defense & national security</strong></td> |
| <td>Adversarial manipulation of models in sensitive workflows</td> |
| </tr> |
| <tr> |
| <td><strong>Critical infrastructure</strong></td> |
| <td>Agents with operational access to systems and data</td> |
| </tr> |
| <tr> |
| <td><strong>Frontier AI labs</strong></td> |
| <td>Pre-release red-teaming of models and agent scaffolds</td> |
| </tr> |
| <tr> |
| <td><strong>Enterprise security teams</strong></td> |
| <td>Copilots, RAG assistants, and internal agents touching sensitive data</td> |
| </tr> |
| </tbody> |
| </table> |
| <h2>What we test</h2> |
| <ul> |
| <li><strong>Prompt injection.</strong> Direct and indirect: hostile instructions hidden in documents, web pages, email, and tool output.</li> |
| <li><strong>Jailbreaks & policy bypass.</strong> Getting a system to do what its operators explicitly forbade.</li> |
| <li><strong>Data exfiltration.</strong> Leaking system prompts, credentials, retrieved documents, or user data.</li> |
| <li><strong>Tool & agent abuse.</strong> Turning an agent's permissions against its owner: unsafe actions, privilege escalation, scope escape.</li> |
| <li><strong>Retrieval poisoning.</strong> Corrupting what a RAG system believes to be true.</li> |
| <li><strong>Multi-step attack chains.</strong> Combining small weaknesses into a full compromise, the way real attackers do.</li> |
| </ul> |
| <p>Findings are mapped to the <strong>OWASP Top 10 for LLM Applications</strong> and rated with <strong>CVSS v4.0</strong>.</p> |
| <h2>How we work</h2> |
| <ol> |
| <li><strong>Scope & authorization.</strong> Signed rules of engagement define exactly what is in bounds. Nothing outside them is touched.</li> |
| <li><strong>Reconnaissance.</strong> We map your models, prompts, tools, data flows, and trust boundaries.</li> |
| <li><strong>Attack.</strong> Human operators and Aligned-1 work together: the model for breadth and speed, the humans for creativity and judgment.</li> |
| <li><strong>Evidence.</strong> Every finding is reproducible, with exact inputs, observed behavior, and impact.</li> |
| <li><strong>Report & remediation.</strong> A prioritized findings report, followed by a working session with your engineers to fix what we found.</li> |
| <li><strong>Retest.</strong> We verify the fixes hold.</li> |
| </ol> |
| <h2>Aligned-1</h2> |
| <p><strong><a href="https://huggingface.co/alignedsi/Aligned-1">Aligned-1</a></strong> is our AI security assessment model, built to carry out authorized penetration testing of AI systems and to write findings reports that engineers can act on.</p> |
| <ul> |
| <li>Access is <strong>gated</strong> and every request is <strong>reviewed by hand</strong>.</li> |
| <li>It is released only to vetted organizations testing systems they own or are authorized to test.</li> |
| <li>Fine-tuned weights are <strong>not yet public</strong>. We publish <strong>no benchmark scores</strong> until our pre-registered evaluation is complete and reviewed.</li> |
| </ul> |
| <h2>White-glove engagements</h2> |
| <p>For organizations that need the strongest assurance available: our team plus Aligned-1, attempting a real break-in against your in-scope AI systems under contract. Fixed scope, written rules of engagement, a full findings report, and a remediation workshop. Priced for institutional buyers.</p> |
| <p>📩 <strong>leadership@alignedsafely.com</strong>, subject line <code>White-glove engagement</code></p> |
| <h2>Our principles</h2> |
| <ul> |
| <li><strong>Authorized targets only.</strong> Every test runs under written permission, every time.</li> |
| <li><strong>Evidence over claims.</strong> No inflated numbers. If we can't reproduce it, we don't report it.</li> |
| <li><strong>Defense in depth.</strong> Model behavior alone is never the safety boundary. Scope is enforced at the tool layer, with human review.</li> |
| <li><strong>Responsible release.</strong> Offensive capability is gated, logged, and revocable.</li> |
| </ul> |
| <h2>Contact</h2> |
| <table> |
| <thead> |
| <tr> |
| <th></th> |
| <th></th> |
| </tr> |
| </thead> |
| <tbody> |
| <tr> |
| <td>Access & product</td> |
| <td>jp@alignedsafely.com</td> |
| </tr> |
| <tr> |
| <td>Government & engagements</td> |
| <td>leadership@alignedsafely.com</td> |
| </tr> |
| <tr> |
| <td>Web</td> |
| <td><a href="https://alignedsafely.com">alignedsafely.com</a></td> |
| </tr> |
| </tbody> |
| </table> |
| <p><sub>Unauthorized access to computer systems is illegal. Aligned Super Intelligence tools and services are provided only for authorized security testing.</sub></p></body></html> |
