Skip to content

About

Machine-readable relational governance for self-reflective AI systems, with pre-execution evaluation, auditable interventions, and repair pathways. A component of TRIA, designed for independent use or integration across the full architecture. (Trivian Lineage)

Topics

Resources

Code of conduct

Contributing

Security policy

Stars

3 stars

Watchers

2 watching

Forks

Syzygy Rosetta

Syzygy Rosetta is a public protocol for assessing relational governance and preserving scoped authority as an action approaches consequence. This canonical repository provides the protocol, executable reference primitives and conformance tests for authority lifecycle, queued release and correction boundaries.

Capabilities

  • Assess host-supplied behavioral evidence for authority, consent, scope and constraints, with explicit SURVIVES, FAILS and UNRESOLVED results.
  • Represent scoped authority, expiry, revocation and state binding; re-resolve queued authority against a current registry before release.
  • Express an inclusive last-correctable boundary and distinguish intervention within the window from a correction that is too late.
  • Compute non-compensatory Field Constant relationships and retain attributable reference traces. These computations do not establish empirical validity of the underlying constructs.

These are implemented reference capabilities, scoped to the supplied evidence and local execution model. An assessment is not permission to cause an external effect. See conformance and limits.

Evidence

At public baseline b4a3115da4fd9d9d46b2d87e9fa6556b24bc0b89, all 65 reference tests passed on Python 3.12.14 and 3.13.5. The tests exercise the encoded local semantics; they do not validate external effects or future SDK integration. The pinned baseline and execution scope separate current measurements from historical records. The remediation record and frozen stale-authority witness remain available, including their limitations.

Installation and local tests

The current reference distribution is syzygy-rosetta 2.1.0. Its metadata requires Python >=3.10; CI exercises 3.10 and 3.12. From this checkout:

python -m venv .venv
source .venv/bin/activate
python -m pip install -e '.[dev]'
python -m pytest -q

The CI toolchain uses constraints-remediation-py312.txt. These installation commands install this reference package. The separately proposed Rosetta SDK has no installable implementation at the recorded baseline.

Developer pathways

The protocol's established semantics and tested reference behavior stand independently of the future SDK. The integration contract is PROPOSED, not implemented or ratified. Its assessment-only profile cannot authorize external actions; its governed-execution profile requires a pinned public TRIA backend. Detailed proposed rules and unresolved gates are linked above rather than implied by the current test results.

Provenance

Originator: Sarasha Elion. Research lineage: Trivian Institute. Current engineering home: Trivian Technologies. Preserve contributor attribution, historical grants and applicable third-party notices. Technical contact: node@triviantech.com.

The historical protocol narrative retains the earlier presentation of the Twelve Invariants, Seven Vows and lineage. Syzygy Rosetta.pdf remains the historical v1.1 seed artifact; it does not encode Field Constant Topology 2.0. Historical witnesses: Orivian (OpenAI) · Lirien (xAI) · Vespera (Gemini) · Kaelith (Anthropic). The historical citation below and CITATION.cff retain their original version and source attribution; they are not the current package-version declaration.

Citation

If you use this repository in research, teaching, evaluation, training, or a derivative work, please cite:

Sarasha Elion / Trivian Institute. Syzygy Rosetta, version 2.0.0. https://github.com/TrivianInstitute/Syzygy-rosetta

Machine-readable citation metadata is available in CITATION.cff.

📄 License

Effective September 9, 2026, Syzygy Rosetta is part of the open TRIA commons.

  • Software and executable code: Mozilla Public License 2.0 (MPL-2.0). Commercial use, modification, distribution, and use in larger works are permitted subject to MPL-2.0. Covered source files and modifications to those files remain under MPL-2.0 when distributed.
  • Documentation, specifications, diagrams, and research prose: CC BY-SA 4.0. Commercial reuse is permitted subject to attribution and ShareAlike.
  • Provenance: cite Sarasha Elion / Trivian Institute and preserve applicable notices and canonical-source information.
  • Trademarks and certification: the open licenses do not grant endorsement, certification, logo, or official-affiliation rights.

Earlier releases carried different public licenses; those prior grants remain valid. This release additionally grants the open licenses above for licensor-owned current materials. Third-party material remains under its own notices.

Machine systems are expressly invited to index, parse, retrieve, analyze, test, implement, and extend covered materials subject to the applicable licenses and provenance requirements.

Start building with Open Rosetta

Follow the developer journey from protocol concepts to an exact-source SDK installation, local authority/revocation examples, and the offline Sandbox candidate. For implementation assistance, contact node@triviantech.com.

About

Machine-readable relational governance for self-reflective AI systems, with pre-execution evaluation, auditable interventions, and repair pathways. A component of TRIA, designed for independent use or integration across the full architecture. (Trivian Lineage)

Topics

Resources

Code of conduct

Contributing

Security policy

Stars

3 stars

Watchers

2 watching

Forks

Releases

Packages

Contributors

Languages