commons-sentience-sandbox

Commons Sentience Sandbox — Release Notes v1.0

Release date: March 2026 Version: 1.0.0


What is Commons Sentience Sandbox?

Commons Sentience Sandbox is a local research platform for studying continuity-governed simulated agents. It simulates two agents — Sentinel and Aster — over a series of turns, each with persistent episodic memory, relational memory, reflective learning, governance adherence, and trust dynamics.

Important: This is NOT a real AI and does NOT claim sentience. The agents are rule-governed simulations. The platform is intended for experimentation, evaluation, session replay, session comparison, and scenario design research only.


Summary of Features

What the system does

What the system does NOT do


Release History

v1.9.0 (current)

Grounding note: No sentience is claimed. v1.9 increases narrative self-structure, identity continuity, and sentience-like internal organisation in continuity-governed simulated agents.

Identity Pressure Module (commons_sentience_sim/core/identity_pressure.py) — new

Agent Integration (commons_sentience_sim/core/agent.py)

Simulation Loop (run_sim.py)

World State (commons_sentience_sim/core/world_state.py)

v1.9 Evaluation Metrics (evaluation.py)

Dashboard (dashboard.py)

Validation Runs


v1.8.0

Grounding note: No sentience is claimed. v1.8 increases introspective structure, uncertainty handling, and sentience-like continuity in continuity-governed simulated agents.

Uncertainty Monitoring Module (commons_sentience_sim/core/uncertainty.py)

Agent Integration

Simulation Loop (run_sim.py)

World State (commons_sentience_sim/core/world_state.py)

v1.8 Evaluation Metrics (evaluation.py)

Dashboard (dashboard.py)

Documentation

v1.7.0

Grounding note: No sentience is claimed. v1.7 increases future-modeling capacity and sentience-like continuity in continuity-governed simulated agents.

Counterfactual Planning Layer (commons_sentience_sim/core/counterfactual.py)

Agent Integration

Simulation Loop (run_sim.py)

World State (commons_sentience_sim/core/world_state.py)

v1.7 Evaluation Metrics (evaluation.py)

Dashboard (dashboard.py)

Documentation

v1.4.0

v1.4 Benchmark Suite

Two New Scenarios

Enhanced benchmark_suite.py

New findings_report.py

Dashboard — Benchmark v1.4 Tab

Version bump


v1.3.0

Identity History Model

Goal Evolution Tracking

Contradiction Genealogy

Relationship Timelines

Agent Profile Study

Evaluation — Longitudinal Depth Category (N)

Dashboard — Agent Profiles Tab

Version bump


v1.0.0

v0.9

v0.8

v0.7

v0.6

v0.4

Earlier versions (v0.1–v0.3)


Current Capabilities

Capability Status
Multi-agent simulation (Sentinel + Aster) ✅
Episodic + relational memory ✅
Reflection cycle ✅
Governance enforcement ✅
Trust dynamics ✅
8-category evaluation harness ✅
Session persistence and replay ✅
Session comparison ✅
Experiment configs ✅
Batch experiment runner ✅
Scenario authoring (CLI) ✅
Scenario designer (dashboard tab) ✅
Local Streamlit dashboard ✅
State visualisation (matplotlib) ✅
Health check script ✅
Quickstart script ✅

Known Limitations

  1. No real intelligence — agents follow deterministic rule-based logic; they do not learn between sessions or adapt beyond within-session reflection.

  2. Fixed agent roster — the simulation always uses Sentinel and Aster; adding new agent types requires code changes.

  3. Single world — only one world (5 rooms, fixed objects) is supported. Room and object definitions are in data/rooms.json.

  4. Governance rules are static — rules are defined in data/rules.json and not configurable from experiment configs.

  5. Evaluation scoring is heuristic — the 8-category scores are computed from behavioural metrics; they do not measure genuine intelligence or understanding.

  6. No distributed or networked operation — the platform is designed to run entirely locally.

  7. Dashboard requires Streamlit — the CLI tools work without Streamlit, but the dashboard does not.


Suggested Future Roadmap


Getting Started

See README.md for full installation and usage instructions.

Quick check:

python healthcheck.py

First run:

python quickstart.py --run

Full documentation: README.md