Skip to content

DariuszNewecki/CORE

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

2,462 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

CORE

Executable constitutional governance for AI-assisted software development.

Designed for environments where AI action traceability is not optional.

License: MIT Release Docs Autonomy

Versioning: SemVer with 2.x denoting an API approaching stability (Beta on PyPI); see the versioning policy.


The Problem

AI coding tools generate code fast. Too fast to stay sane.

Without enforcement, AI-assisted codebases accumulate invisible debt — layer violations, broken architectural contracts, files that grow unbounded. And agents, left unconstrained, will eventually do something like this:

Agent: "I'll delete the production database to fix this bug"
System: Executes.
You:    😱

CORE makes that class of violation impossible — structurally blocked before execution, not detected after the fact. (Which surfaces hard-block versus advisory-report is mapped under Current proof status below.)

Agent: "I'll delete the production database to fix this bug"
Constitution: BLOCKED — Violates data.ssot.database_primacy
System: Execution halted. Violation logged.
You:    😌

CORE is a governance runtime that constrains AI agents with machine-enforced constitutional law — enforcing architectural invariants, blocking invalid mutations automatically, and making autonomous workflows auditable and deterministic.

LLMs operate inside CORE. Never above it.


🎬 See it govern itself — one command

You don't have to take this on faith. On a clean machine with Docker:

git clone https://github.com/DariuszNewecki/CORE.git
cd CORE && ./install-core.sh

install-core.sh stands up CORE and finishes by running the consequence-chain demo live — no LLM key required. It:

  • commits a function that violates linkage.assign_ids (a blocking rule)
  • watches CORE's audit block it
  • has CORE propose a fix, the governor approve it, and CORE execute it and commit the repair
  • re-audits to confirm clean
  • prints the full causal chain it recorded: finding → proposal → approval → execution → file change

Re-run it any time with scripts/demo.sh.

Governance is executable.


📋 Live Audit Trail

Every enforced action records its lineage. Two consequence chains, pulled live from the CORE database — same schema, two different authorities:

Autonomous path — risk-classified as safe, system self-approved

FINDING     → workflow.ruff_format_check       src/api/cli/client.py                2026-05-18 05:15:15 UTC
PROPOSAL    → 8845dacc…   fix.format                                                2026-05-18 05:16:15 UTC
APPROVAL    → risk_classification.safe_auto_approval                                2026-05-18 05:16:15 UTC
EXECUTION   → completed   (1.29s)                                                   2026-05-18 05:17:18 UTC
FILE CHANGE → +105 / -0   98da9038 → fca9a971  src/api/cli/client.py                2026-05-18 05:17:19 UTC

Human-approval path — governor in the loop

FINDING     → purity.docstrings.required       src/cli/commands/audit_reporter.py   2026-05-15 08:28:29 UTC
PROPOSAL    → a4363a81…   fix.docstrings                                            2026-05-16 13:39:34 UTC
APPROVAL    → principal.governor  (cli_admin)                                       2026-05-16 13:53:32 UTC
EXECUTION   → completed   (24.5s)                                                   2026-05-16 13:55:48 UTC
FILE CHANGE → +26 / -0    5a123426 → 71fde489  src/cli/commands/audit_reporter.py   2026-05-16 13:55:49 UTC

Both chains are queryable end-to-end from proposal_consequences and blackboard_entries. The constitution decides which authority applies; the schema is identical. Reproduce them yourself with the Consequence-chain query in the Proof Index.


Architectural Model

CORE separates responsibility across four repository layers — three enforced as constitutional law, and Specs (human intent). This separation is enforced as law — not convention.

📐 Specs — Human Intent (.specs/)

Where humans define what the system is for and why decisions were made. Contains architectural papers, northstar documents, user requirements, architectural decision records, and planning documents. This is the entry point for anyone trying to understand CORE before reading its implementation.

.specs/ is read by humans and searchable by CORE's semantic layer. It is never written by CORE itself.

🧠 Mind — Law (.intent/ + src/mind/)

Defines what is allowed, required, or forbidden. Contains machine-readable constitutional rules, enforcement mappings, phase-aware governance models, and the authority hierarchy (Meta → Constitution → Policy → Code).

Mind never executes. Mind never mutates. Mind defines law.

⚖️ Will — Judgment (src/will/)

Reads constitutional constraints, orchestrates autonomous reasoning, and records every decision with a traceable audit trail. Every operation follows a structured phase pipeline:

INTERPRET → PLAN → GENERATE → VALIDATE → STYLE CHECK → EXECUTE

Will never bypasses Body. Will never rewrites Mind.

🏗️ Body — Execution (src/body/)

Deterministic, atomic components: analyzers, evaluators, file operations, git services, test runners, CLI commands.

Body performs mutations. Body does not judge. Body does not govern.


How CORE Works

Every autonomous operation is governed by the same constitutional loop:

flowchart TD
    A["🟢 GOAL\nHUMAN INTENT"] --> B["📂 CONTEXT\nRepo state • knowledge • history"]
    B --> C["🔒 CONSTRAINTS\nImmutable rules\n255 rules • 15 engines"]
    C --> D["🗺️ PLAN\nStep-by-step reasoning\nRule-aware plan"]
    D --> E["✨ GENERATE\nCode • changes • tool calls"]
    E --> F["✅ VALIDATE\nDeterministic checks\nAST • semantic • intent • style"]
    F -->|Pass| G["▶️ EXECUTE\nApply compliant changes"]
    F -->|Fail| H["🔄 REMEDIATE\nRepair violation\nAutonomy Ladder"]
    H --> E
    G --> I["✓ SUCCESS\nChanges committed"]

    subgraph "SAFETY HALT"
        direction TB
        J["🚨 CONSTITUTIONAL VIOLATION\n→ HARD HALT\n+ FULL AUDIT LOG"]
    end

    E -.->|Any violation| J
    F -.->|Any violation| J

    classDef phase      fill:#f8f9fa,stroke:#495057,stroke-width:2px
    classDef constraint fill:#d1e7ff,stroke:#0d6efd,stroke-width:2.5px
    classDef validate   fill:#fff3cd,stroke:#ffc107,stroke-width:2.5px
    classDef halt       fill:#ffebee,stroke:#dc3545,stroke-width:3px

    class A,B,D,E,G,I phase
    class C constraint
    class F validate
    class J halt
Loading

System Guarantees

Within CORE:

  • No file outside an autonomy lane can be modified
  • No structural rule can be bypassed silently
  • No atomic action can execute outside the governed executor (inline authorization is deferred to the audit→consequence loop)
  • Decisions are phase-aware and logged with decision traces (audit persistence is best-effort — see Current proof status)
  • No agent can amend constitutional law

If a blocking rule fails, execution halts with no partial state. Reporting and advisory rules surface findings and continue — what blocks versus what reports depends on the mode.

Current proof status

CORE's guarantee semantics are split across modes by design. This is the honest map of what each surface does, so a single binary claim ("CORE blocks violations") is not mistaken for the whole picture:

Surface Mode Behaviour
.intent/ writes hard invariant blocked — the governance directory is immutable to all components
Constitutional rules always-blocking block a commit regardless of strict mode
Policy rules strict vs. default block only when strict_mode=True; otherwise report
Capability tier advisory today reports a "would-deny" signal; does not yet block (ADR-079)
Stateless CI (GitHub Action) rule subset skips knowledge_gate + llm_gate (they need DB / LLM state) and reports the skip
Action audit trail best-effort recorded when the DB write succeeds; a write-action failure is surfaced (AUDIT_GAP), not silent

The hard invariants and constitutional rules block unconditionally; the policy, capability, and stateless tiers are weaker by design and labelled here so the boundary is legible rather than implied.


Constitutional Primitives

Primitive Purpose
Document Persisted, validated artifact
Rule Atomic normative statement
Phase When the rule is evaluated
Authority Who may define or amend it

Enforcement strengths: Blocking · Reporting · Advisory


Enforcement Engines

Engine Method
ast_gate Deterministic structural analysis (AST)
regex_gate Pattern-based text enforcement
glob_gate Path and boundary enforcement
cli_gate CLI surface and command-shape enforcement
artifact_gate Declared-vs-discovered artifact completeness
workflow_gate Phase-sequencing and coverage checks
knowledge_gate Responsibility and ownership validation
action_gate Atomic-action invariants
passive_gate Substrate-enforced rules (DB/runtime marker)
taxonomy_gate Capability-id ↔ atomic-action coherence (ADR-079 D9)
contracts_gate Cross-cutting data-contract coherence (context-level; ADR-102)
llm_gate LLM-assisted semantic checks
IntentGuard* Runtime write authorization (not audit)

*Runtime Gate per .specs/papers/CORE-Gate.md, kept here for visibility.

Deterministic when possible. LLM only when necessary.

255 rules across 58 rule documents. 247 are mapped to enforcement engines; the other 8 are advisory rules that are mapping-exempt by design (#820), so every rule that requires a mapping has one. "Mapped" means engine-bound — not enforced in every mode: stateless CI skips knowledge_gate and llm_gate, which need the knowledge graph and an LLM provider.


The Autonomy Ladder

CORE progresses through defined levels. Each adds capability while remaining constitutionally bounded.

A0 — Self-Awareness       ✅  Knows what it is and where it lives
A1 — Self-Healing         ✅  Fixes known structural issues automatically
A2 — Governed Generation  ✅  Natural language → constitutionally aligned code
A3 — Governed Autonomy    ✅  Daemon finds, proposes, and fixes violations unattended  ← current
A4 — Self-Replication     🔮  Writes CORE.NG from its own understanding of itself

Production Readiness

CORE does not claim a status it has not earned. The gates are defined in URS-production-readiness; the current verdict and per-gate evidence live in the attestation manifest. This table is generated from that manifest — do not edit it by hand.

Production readiness: NOT ATTESTED. All gates must hold simultaneously, and each met gate must carry dated, human-signed evidence. There is no composite score.

Progress: 0/15 met · 11 partial · 1 not met · 1 not demonstrated · 2 not started.

Gate Status Primary gap
G1 — Fresh-install proof ⚠️ partial CI-automated fresh-install absent; manual cold-room only
G2 — Constitutional enforcement is deterministic and regression-tested ⚠️ partial Per-rule known-violating + known-compliant fixtures not universally applied
G3 — Layer integrity is machine-enforced and blocking ⚠️ partial Several layer rules at reporting severity (no_body_to_will, will.no_direct_database_access); reporting-only set not formally documented
G4 — Autonomous loop reliability demonstrated by soak ❌ not demonstrated No 72h soak run; two loop bugs fixed 2026-07-05
G5 — Mutation-lane equivalence and lifecycle safety ❌ not met RemediationCeremony(write=True) bypasses the governed lifecycle (#818); claim.proposal path unaudited this cycle; approval_authority population unverified
G6 — Risk model is governed, correct, and regression-tested ⚠️ partial else-branch safety net exists for future StepKinds; known set handled but exhaustive-handling not independently audited
G7 — Circuit breakers operate at the correct granularity ⚠️ partial Flow-level and worker-level circuit breakers not formally audited
G8 — Integration tests for the governed mutation chain ⬜ not started No end-to-end integration test of the governed mutation chain (cognitive delegate -> write -> sandbox -> evidence). Integration INFRASTRUCTURE exists — the specific chain test does not.
G9 — Enforcement integrity fails closed ⚠️ partial Unmapped non-advisory rules can still yield PASS (#822); empty-graph vacuous-pass guard not built
G10 — Operator observability ⚠️ partial Questions 4 (failure diagnosis) and 5 (rollback) not confirmed answerable by a non-author without source access
G11 — Upgrade and migration safety ⬜ not started Schema-as-dump (db_schema_live.sql); core._migrations records only (id, applied_at) — no delta path, no version sequence. Most critical structural gap.
G12 — Runtime trust boundary is audited and documented ⚠️ partial Rate-limiting posture not fully documented; user-facing route safety not fully confirmed per-route
G13 — Documentation for operators ⚠️ partial No operator runbook; CLAUDE.md is comprehensive for Claude Code but is not an operator document
G14 — Source reversibility ⚠️ partial Source-restoration-after-completed-mutation not proven by test; rollback-vs-forward audit distinction not verified
G15 — Release integrity ⚠️ partial Release-gating checks not fully enforced as a blocking pre-publish gate; release provenance (tag/SHA/check-run) not queryable

The full gate definitions and acceptance criteria are in URS-production-readiness.md; the evidence and blockers behind each status are in production-readiness.yaml.


Requirements

Dependency Version
Python 3.12+
PostgreSQL ≥ 14
Qdrant latest
Docker for services
Poetry for deps

Quick Start

Honest status — what works today. CORE governs itself end to end (the demo above), and audits any repo that has a .intent/ constitution — in CI via the GitHub Action, or locally with core-admin code audit --offline inside that repo. pip install core-runtime gives you the core-admin CLI.

Govern your own repo (BYOR): pip install core-runtime and two commands bootstrap a fitted constitution. (1) core-admin project onboard <path> --write delivers the machinery floor (no LLM, no database). (2) core-admin project scout <path> --write reads your source, proposes candidate rules via LLM, and requires you to ratify each before delivery — no LLM means a curated four-rule menu instead. core-admin code audit --offline inside that repo enforces the ratified rules immediately. The machinery floor is bundled in the wheel — both commands work from a plain pip install core-runtime.

Fastest way to see CORE today: run it on itself, below.

Full local runtime — one command. Clone, then run the installer:

git clone https://github.com/DariuszNewecki/CORE.git
cd CORE
./install-core.sh

install-core.sh checks prerequisites, installs dependencies, starts Postgres + Qdrant, applies the schema, and finishes by showing CORE govern itself — a violation found, proposed, approved, fixed, and verified, with the consequence chain recorded. No LLM API key needed for the demo.

Prefer to run the steps yourself?
poetry install
cp .env.example .env
docker compose up -d
docker compose exec -T postgres psql -U postgres -d core < infra/sql/db_schema_live.sql
poetry run core-admin code audit --offline   # offline mode needs no running services

Full setup — services, schema, vector sync, first audit — is in Getting Started.


Documentation

Full documentation, architecture deep-dive, and governance reference: dariusznewecki.github.io/CORE

To understand what CORE is for before reading its implementation, start here: .specs/northstar/CORE-What-It-Does.md


Issue Tracker Guide

CORE's tracker mixes governance-internal bookkeeping with ordinary engineering work, so labels do double duty. type:*, priority:*, status:*, band:*, and the *-debt labels are CORE's own governed taxonomy — they say what kind of work an issue is and where it sits relative to the roadmap (roadmap marks epics on a named milestone; technical-debt/governance-debt mark known gaps CORE is carrying forward deliberately, not bugs; band:E marks outward-facing/adoption work). Two standard labels tell you where to actually start: good-first-issue is scoped enough to attempt without deep repo context, and help-wanted is open to outside contribution but assumes you've read the ADR(s) it links. If an issue carries neither — and especially if it carries type:question — it's a live design question awaiting a maintainer decision, not something to open a PR against yet.


Project Status

Current Release: v2.9.1 — CLI Client Parity

Active work: A3 Governed Autonomy — the daemon runs continuously, finds constitutional violations in its own codebase, proposes fixes, executes approved fixes, and verifies the result. The governor's role is to define intent, review proposals that require architectural judgment, and approve constitutional changes.

All four A3 integrity gates are now closed. No enforcement logic or operational threshold lives in src/ — governance is declared in .intent/ and enforced from there. The autonomous loop is circuit-breaker protected; systematic errors surface as signals rather than unbounded churn.

Gate Meaning Status
G1 — Loop closure Round-trip autonomous fix demonstrated
G2 — Convergence Circuit-breaker; resolution rate > creation rate
G3 — Consequence chain Causality queryable end-to-end
G4 — Governance in .intent/ No enforcement logic or thresholds in src/

License

License: MIT


Build fast with AI. Stay constitutionally aligned.

About

CORE is a governance runtime for autonomous AI systems. It enforces constitutional rules during execution, prevents governance bypass, and creates auditable authority chains for agent actions across operational domains.

Topics

Resources

Contributing

Security policy

Stars

Watchers

Forks

Releases

Sponsor this project

Packages

Used by

Contributors

Languages