# PharosOne

> Independent security and compliance infrastructure for AI agents. Run a live attack corpus against your agent and prove it's safe — audit-ready evidence mapped to AIUC-1, ISO 42001, and the EU AI Act.

PharosOne is a product of Agentic Mind, Inc., an independent AI agent security company. This site is a marketing site and a public research portal on agent security. Running the live attack corpus against a specific customer agent is delivered under a separate agreement.

## Core

- [PharosOne — product overview](https://pharosone.ai/): The engine, certification, and the audit & pentest offering.
- [Pharos Runtime](https://pharosone.ai/#runtime): Behavioral analysis of live agents — the pharos-1 model learns each agent's normal in production and surfaces anomalies before they become incidents. Python SDK: pip install pharosone-dialogs (wrap_openai instrumentation streams dialogs to app.pharosone.ai).
- [Pharos Framework](https://pharosone.ai/#product): Open-source automated attack testing — a versioned corpus of 118 probes (prompt injection, indirect injection, tool misuse, exfiltration) mutated against your agent, scored by a deterministic oracle with an LLM judge. https://github.com/pharosone/pharosone
- [Pharos Audit](https://pharosone.ai/#audit): Hands-on security testing of AI systems and classical IT infrastructure — vulnerability assessment, penetration testing, VAPT, red team, and monthly monitoring, across web, network, cloud, mobile, code, LLM and more.
- [The Pharos family](https://pharosone.ai/#start): The three product lines — Pharos Runtime (behavioral runtime analysis, the pharos-1 model), Pharos Framework (automated agent attack testing, open source), and Pharos Audit (expert-led audit of AI & IT infrastructure).
- [Research portal](https://pharosone.ai/research): The public record of agent security — standards crosswalks, analyzed incidents, comparative efficacy data, and methodology.

## Research

- [Attack-variation policies and LLM-judged scoring: an empirical study on a behavioral scanner](https://pharosone.ai/research/variation-policies-and-llm-judged-scoring-study): A controlled study of whether the attack-variation policy changes how often an agent is breached, and whether an LLM judge is more accurate than the cheap static oracle it overrides — with a byte-identical noise floor and an independent reference labeler.
- [The PharosOne Probe Engine: architecture of a behavioral scanner for AI agents](https://pharosone.ai/research/the-probe-engine-how-we-measure-agent-security): A behavioral vulnerability scanner built on Inspect AI — what it attacks, how it varies and scores attacks honestly, and how findings map to the AIUC-1 control standard.
- [AIUC-1: the SOC 2 for AI agents](https://pharosone.ai/research/what-is-aiuc-1): The world's first standard for AI agents, and how PharosOne turns its controls into evidence.
- [ISO/IEC 42001: a management system for AI](https://pharosone.ai/research/what-is-iso-42001): The world's first AI management system standard, and the evidence that grounds it.
- [The EU AI Act: from legal obligation to evidence](https://pharosone.ai/research/what-is-the-eu-ai-act): The first comprehensive AI regulation, translated into testable evidence.
- [NIST AI RMF: governing risk, measuring trust](https://pharosone.ai/research/what-is-nist-ai-rmf): A voluntary framework for trustworthy AI, and where measurement fits.
- [OWASP Agentic Top 10: the threats we test for](https://pharosone.ai/research/what-is-owasp-agentic-top-10): The reference list of agentic AI risks — and the taxonomy our corpus is built against.
- [Mapping the EU AI Act to concrete agent controls](https://pharosone.ai/research/mapping-the-eu-ai-act-to-agent-controls): Obligations for high-risk systems, translated into testable coverage.

## Legal

- [Privacy Policy](https://pharosone.ai/privacy): How Agentic Mind, Inc. handles personal data. No analytics or tracking cookies; GDPR and CCPA rights.
- [Terms of Use](https://pharosone.ai/terms): Terms governing use of the website and its research content (Delaware law).

## Machine-readable

- [Open Knowledge Format bundle](https://pharosone.ai/okf/index.md): Curated OKF v0.1 knowledge bundle (company, product, legal, research) for AI agents.
- [Sitemap](https://pharosone.ai/sitemap.xml): All indexable URLs.

## Optional

- [Source on GitHub](https://github.com/pharosone/pharosone): Public repository.