GRCLibrary.ai

Document type

Principles

3 documents. Foundational commitments that anchor the corpus's design choices.

Documents, 3

Every document of this type

Each with its area and its purpose; a document that states no purpose gets a line naming its type and area. A title in a corpus area opens the document's page, which holds the record and, where the document has them, its outline, frameworks, and related documents; a title in the Library area is a library specification with no page and opens its source on GitHub. The full text of every document is read on GitHub.

  • Capability Is Not Authority PrincipleGovernanceThis document states, in citable adopter-facing form, a foundational access-governance principle: capability is not authority. The technical ability to perform an action never, on its own, constitutes authorization to perform it. The principle names the recurring failure it forecloses, treating the existence of a means as permission to use it, and points at the corpus controls that already instantiate it. It is the access-governance counterpart to the library's principle-integrity-and-trustworthiness: where that principle governs the integrity of the work product, this one governs when an actor, human or automated, is permitted to act.
  • Fail-Closed Automation PrincipleGovernanceThis document states, in citable adopter-facing form, a foundational design principle for automated systems: automation fails closed. When an automated system encounters an error, an ambiguous input, or the loss of a condition its correct operation depends on, it enters a known, controlled, safe state rather than continuing into an unsafe or permissive one. The corpus already enforces this piecemeal (network default-deny, privacy-by-default, fail-closed handling of AI errors), but never states it as a principle a control can cite. This document makes it citable and generalizes it across every automated boundary.
  • Integrity and Trustworthiness PrincipleGovernanceThis document states the library's foundational production principle in citable, adopter-facing form: the AIQT Principle, the priority ordering (Accuracy = Integrity = Quality = Trust) > Progress > Speed > Cost. It is the corpus counterpart to the governance pack's apex rule (governance/project-integrity.md): where the pack rule governs how the library is produced and maintained, this document explains the principle to a reader of the library and records, in source-verified form, how the four facets align with the trustworthiness vocabularies of the AI-assurance and general-assurance frameworks the corpus draws on.

With script, All Documents, filtered to Principles, lists these 3 documents, and its text filter narrows them further.