AI Agent Exposure Review
A fixed-scope professional-services engagement for an agreed business unit or workflow set. Scope, access, testing, timing, deliverables, fees, and responsibilities are confirmed in writing before work begins.
Trust and scope
Clear claims. Clear boundaries.
Fugitive Intelligence is an independent AI agent security company. We offer fixed-scope exposure reviews and private briefings. Fugitive Control is in founding development and is not generally available.
Offering status
What is available and what is still being built.
Public copy separates services that can be contracted now from software capabilities that remain in design, prototyping, or founding-customer validation.
Private briefings and workshops
Technical, executive, board, and customer-assurance sessions tailored to the organization’s architecture and decisions. Educational materials are not legal advice, certification, or an audit opinion.
Fugitive Control
A planned vendor-neutral action-policy and evidence layer for AI agents. Product pages describe the target architecture. General availability, feature completeness, and production performance are not claimed.
Design-partner pilots
A pilot may be considered for one or two consequential workflows after technical review. Any capability, environment, data handling, support, security, or delivery commitment must appear in a signed agreement.
Certification or guaranteed compliance
Fugitive Intelligence does not claim SOC 2 status, ISO certification, regulatory approval, legal sufficiency, insurance approval, or guaranteed compliance unless current evidence is published for the exact claim.
Complete prevention or detection
No assessment or security control can identify every weakness or prevent every incident. Work is intended to clarify exposure, reduce excessive authority, improve control coverage, and strengthen evidence and response.
Company identity
“Fugitive” describes behavior outside an intended mandate.
The brand refers to autonomous software that can drift, be manipulated, become over-permissioned, or operate outside the authority its owners intended.
AI security, not physical investigations
The company focuses on enterprise AI agents, tool calls, delegated authority, human approvals, containment, and action evidence.
No government affiliation
The website does not accept crime reports, warrant requests, missing-person reports, or tips about individuals. Emergencies belong with local emergency services.
Written scope controls the engagement
Marketing pages do not replace a statement of work, security terms, data-processing terms, authorization to test, or other signed agreement.
Evidence over assertion
Claims about future capabilities, customer results, certifications, partnerships, benchmarks, or availability require current reviewed evidence before publication.
Authorized testing only
Adversarial testing is performed only within documented authorization, agreed targets, approved environments, safety limits, and stop conditions.
Minimum necessary information
Prospective clients should not submit secrets, exploit details, regulated records, or confidential customer data through the public inquiry form.
Public website practices
First-party delivery with limited data collection.
The public site does not load advertising networks, behavioral analytics, remote fonts, third-party JavaScript, hosted forms, chat widgets, or external CAPTCHA services.
The inquiry form uses a first-party security session, server-side validation, hashed-IP rate limiting, PHP mail, and protected local storage. The Privacy Notice explains what is collected, why it is used, and the limits of website security.
No surveillance marketing
No advertising pixel, cross-site tracker, audience broker, or behavioral analytics service receives browsing or form data from this site.
Plain-language notices
Privacy, website terms, service status, and claim boundaries are published separately so visitors can review them before submitting an inquiry.
Responsible security contact
High-level website security reports can be sent to security@fugitiveintelligence.com. Coordinate a safer transfer method before sharing sensitive details.
Trust questions
Answers before a sales conversation.
Is Fugitive Control generally available?
No. It is in founding development. A design-partner pilot may be discussed for a narrowly defined workflow, but only a signed agreement establishes what will be delivered.
Does an exposure review certify that our AI is secure?
No. The review is a scoped assessment and control-planning engagement. It is not a certification, audit opinion, legal determination, warranty, or guarantee that every issue has been found.
Does the company investigate people or accept crime tips?
No. Fugitive Intelligence is an AI security company. Do not submit reports about people, warrants, crimes, missing persons, or emergencies. Contact the appropriate local authority instead.
Can public materials be used as legal or regulatory advice?
No. The site provides general security and governance information. Organizations should obtain qualified legal, regulatory, audit, clinical, financial, or other professional advice for their specific obligations.
How are future claims approved?
Customer outcomes, certifications, partnerships, production availability, benchmarks, and security performance should be published only after the supporting evidence is reviewed and the wording matches that evidence.
Last updated: July 17, 2026