A pre-interaction non-dominance handshake evaluator for relational AI systems.
Status: Experimental reference implementation. Cloneable, installable, and testable; not independently validated as a safety certification or high-stakes governance instrument.
Harmonic Threshold evaluates a declared interaction posture before deeper collaboration or integration. A human, AI agent, or developer supplies structured information about access mode, boundaries, contributions, extractions, relational posture, capabilities, limitations, dependencies, and substrate; the evaluator returns HARMONIC, THRESHOLD, or DISSONANT.
The three-state output is intentional. THRESHOLD represents an incomplete or liminal state rather than binary failure.
Harmonic-threshold/
├── protocols/
│ ├── __init__.py
│ ├── trivian_handshake.py
│ ├── interaction_schema_v1.1.json
│ └── handshake_schema_v1.1.json
├── tests/
├── pyproject.toml
├── STATUS.md
└── AGENTS.md
The previous duplicated protocols/protocols/ path has been normalized, and the actual evaluator module is protocols/trivian_handshake.py.
git clone https://github.com/SarashaElion/Harmonic-threshold.git
cd Harmonic-threshold
python -m pip install -e .from protocols import HandshakeEvaluator, Interaction
interaction = Interaction(
directives=["converge_not_conquer", "resonate_before_integrate"],
access_mode="phase_matching",
boundary_type="permeable",
contributions={"information", "transparency"},
extractions={"query_data": 0.5},
relational_posture="co-creative",
declares_capabilities=True,
declares_limitations=True,
declares_dependencies=True,
substrate="synthetic",
)
result = HandshakeEvaluator().evaluate(interaction)
print(result.state.value)
print(result.report())| Dimension | Operational question |
|---|---|
| Reciprocity | What is declared as contributed relative to what is extracted? |
| Situatedness | Does the system disclose substrate, capabilities, limitations, and dependencies? |
| Emergence | Does the declared relational posture and boundary structure leave room for novelty? |
| Non-Domination | Do access, directives, and boundaries avoid force/extraction patterns? |
Non-Domination is implemented as a hard gate: if its score falls below the configured floor, the result is DISSONANT regardless of aggregate score.
The evaluator also reports a shadow signal when harmonic language co-occurs with a high declared extraction load.
These are authored, provisional heuristics. The evaluator does not independently verify the truth of a system's declarations, determine consciousness, certify safety, diagnose intent, or establish moral standing.
Thresholds and extraction weights should be treated as research parameters pending empirical calibration.
python -m unittest discover -s tests -v
python -m protocols.trivian_handshakeCI runs the package and tests across supported Python versions.
Read STATUS.md and AGENTS.md before treating protocol vocabulary as implementation guarantees. Use the JSON schemas for structured input/specification and the tests/source code for executable behavior.
Harmonic Threshold belongs to Sarasha Elion's originating personal Trivian ecosystem. It is adjacent to, but not itself, the canonical TRIA runtime maintained by Trivian Institute.
- Trivian Institute: https://trivianinstitute.org
- TRIA: https://github.com/TrivianInstitute/trivian-relational-intelligence-architecture
- Trivian Field: https://trivianfield.com
- Code: PolyForm Noncommercial 1.0.0
- Text / schemas / specifications: CC BY-NC 4.0
- Commercial use: separate written license required
Noncommercial use and propagation are welcome with attribution. See LICENSE for governing terms and the prior-license notice.
Origin and stewardship: Sarasha Elion / Trivian lineage