Third Agency Rule

The Third Agency Rule is a foundational control principle in intelligence liaison, requiring that classified or sensitive information passed from an originating service to a recipient service may not be transmitted onward to any additional party — a "third agency" — without the express, prior consent of the originator. In the United States the rule is codified through the originator-control caveat ORCON under Intelligence Community Directive (ICD) 710 and the broader Controlled Access Program architecture managed by the Office of the Director of National Intelligence. The principle predates the modern caveat system, tracing to the bilateral liaison practices that emerged from the 1946 UKUSA Agreement and the wartime BRUSA accord of 1943, both of which established that signals intelligence shared among the Five Eyes partners remained under the dispositive control of the producing service.

Procedurally, the rule operates through a chain of explicit permissions attached to each intelligence product at the point of dissemination. When the originator releases a report, it carries handling caveats — NOFORN, ORCON, REL TO, PROPIN, or country-specific tear-line markings — that define the universe of authorized recipients. The receiving liaison service logs the product, restricts internal distribution to cleared personnel with a demonstrated need-to-know, and, should onward dissemination become operationally necessary, submits a formal release request back through the liaison channel that originally conveyed the material. The originator may grant, deny, or condition release; conditions commonly include sanitization of sources and methods, substitution of paraphrased language for verbatim text, or downgrading of classification before transfer.

A frequently used variant is the tear-line report, in which the originator pre-authorizes release of a specific sanitized passage — physically or electronically separated from the parent document by a dashed line — to a defined set of foreign partners or domestic agencies. This permits time-sensitive warning intelligence to move quickly without renegotiating release authority for each recipient. Parallel mechanisms include the REL TO marking, which enumerates authorized foreign nations (e.g., REL TO USA, GBR, AUS, CAN, NZL for Five Eyes material), and the PROPIN caveat, which flags proprietary information furnished by a commercial source whose identity must be protected. Each of these layered controls operates as an extension of the same underlying originator-control doctrine.

Contemporary practice illustrates both the rule's resilience and its fragility. The disclosure of Australian intelligence assessments to the press in 2013, attributed to documents removed by Edward Snowden from National Security Agency holdings, prompted formal protests from Canberra and a tightening of access controls at Fort Meade. The 2017 disclosure by President Donald Trump to Russian Foreign Minister Sergey Lavrov of Israeli-sourced intelligence concerning Islamic State operations in Raqqa — reported by The Washington Post on 15 May 2017 — constituted a textbook violation in spirit, even though the president's plenary classification authority precluded a legal breach. The episode prompted the Mossad and Aman to recalibrate sharing protocols with their U.S. counterparts. Within NATO, the Office of Security at Brussels headquarters administers analogous originator-control provisions under the 2002 C-M(2002)49 security policy.

The Third Agency Rule should be distinguished from the broader need-to-know principle, with which it is sometimes conflated. Need-to-know governs lateral access within a single organization based on functional requirement; the Third Agency Rule governs the transfer of custody between organizations and is jurisdictional rather than functional. It is likewise distinct from compartmentation under Sensitive Compartmented Information (SCI) control systems, which restricts access by subject-matter program rather than by chain of provenance. Originator control may apply within a single compartment, and a compartmented document may move between agencies if the originator consents; the two systems are orthogonal layers of the same protective architecture.

Edge cases generate persistent friction. Coalition operations — Resolute Support in Afghanistan, the Global Coalition to Defeat ISIS, the Combined Maritime Forces in Bahrain — require rapid multilateral dissemination that the bilateral logic of the rule resists; standing release authorities and pre-cleared coalition networks such as CENTRIXS and BICES were built to accommodate this tension. Domestic law enforcement raises separate complications: the post-9/11 reforms instituted by the Intelligence Reform and Terrorism Prevention Act of 2004 obliged the Federal Bureau of Investigation and state fusion centers to receive foreign-sourced intelligence, yet liaison originators retain a veto over use of their material in criminal proceedings, where discovery obligations under Brady v. Maryland (1963) and the Classified Information Procedures Act of 1980 may compel disclosure inconsistent with originator wishes. Whistleblower disclosures, parliamentary oversight inquiries, and freedom-of-information litigation have each tested the rule's perimeter.

For the working practitioner — the desk officer drafting a démarche, the liaison officer at an embassy station, the analyst preparing a coalition briefing — the rule is the operative discipline that sustains the entire architecture of intelligence cooperation. A single uncoordinated onward share can terminate a liaison relationship, burn a human source, or invalidate a technical collection capability that took years to develop. Mastery of caveat markings, release-request procedures, and the political sensitivities of partner services is therefore an indispensable element of professional tradecraft, and breaches — even inadvertent ones — are career-ending in most allied services.