General Reasoning publishes open standards that define the governance and compliance framework for the Chandra ecosystem. All standards are versioned through the Chandra chain — the chain is the version history.
The open protocol specification for append-only, hash-chained, attributed audit records. Defines the context unit structure, chain integrity rules, attribution requirements, snapshot format, and tickler mechanism. MIT licensed.
Governed AI Boundary Attestation. Defines the formal framework for AI inference boundary risk acceptance. Every AI inference boundary crossing in a governed deployment requires a Chandra context unit. The GABA cert is the deployment-level attestation that the framework is in place.
Consolidate, Reduce, Close. The minimum surface standard for AI boundary governance. Defines a scoring model across three pillars: Boundary, Attribution, and Halt. A deployment operating with Chandra as its audit infrastructure achieves Boundary pillar score 4 by construction.
Chandra Audit Maturity Model. Defines maturity levels for organizations adopting Chandra as their audit infrastructure. Level 1 through Level 5, from basic CU production through full forest federation with cross-instance trust.
Whitepaper available at chandraprotocol.com.
Auditable by Construction — The foundational paper. The audit record as gate token, not receipt.
Arbitrary Audit Query — How SPARQL over AllegroGraph enables any cross-spoke compliance query.
Auditable Authorization — RBAC and identity governance as a Chandra-native construct.
Energy-Compute Renewal Thesis — Chandra as the governed infrastructure for distributed energy and compute markets.
Personality Layer Specification — The vocabulary mapping framework that renders Chandra primitives in domain-specific regulatory terminology.
Cross-Forest Trust Protocol — Governed federation between independent Chandra instances with cryptographic trust anchors.
The Ledger Standard — Chandra as the compliance infrastructure for any governed exchange of value between identified participants.