Cubiczan/trust-ledger-os

Trust Ledger OS is a trust and risk control plane for AI teams. Every high-impact change is reviewed, traced, and recorded before it reaches customers or cash.
Product screenshots
Trust Ledger OS desktop overview explaining the review workflow and its supporting tool stack.
Existing UI capture stored in this repository; displayed values may be demo or sample data.
Overview
The repo combines a public landing page with four scaffolded implementation phases. The same roadmap is mirrored into reusable package boundaries so the app, docs, and future library work stay aligned.
Mirrors:
- icohangar-ops/trust-ledger-os
- Cubiczan/trust-ledger-os
Core Roles
- PRISM: runtime observability for agent calls, approvals, and latency.
- Prelint: pre-merge product review for policy drift, bad defaults, and risky logic.
- GIDE: secure offline editing and emergency fixes when the network is not trusted.
What It Does
1. Capture a code change, spend request, or agent action.
2. Review it against product policy and trust rules.
3. Attach a PRISM trace to the runtime decision.
4. Produce an approve, deny, or counter outcome.
5. Store the decision in a ledger future runs can reuse.
Live Routes
- / landing page and roadmap
- /foundation phase 1 scaffold
- /framework-benchmark phase 2 scaffold
- /research-reasoning phase 3 scaffold
- /production-controls phase 4 scaffold
- /decisions decision ledger (CHP decision substrate) · GET /api/decisions live read
Phase Packages
- foundation-kit/: shared agent glossary, decision tree, model/tool matrix, and prompt skeletons.
- framework-benchmark-suite/: framework profiles, comparison cases, and scoring rubric.
- research-reasoning-kit/: research pipeline stages and reusable reasoning patterns.
- production-controls-kit/: eval harness, drift monitoring, guardrails, and release gate scaffold.
- decision-substrate/: CHP decision records — R0 gate, deterministic foundation scoring (guardrails 40 / bounded result 30 / parity 30), configurable per-domain floors (general 70, finance 100, blockchain/DeFi 85), substrate-enforced human locks (TRUST_LEDGER_CHP_REQUIRE_HUMAN_LOCK, default on), and an append-only JSONL ledger sealed with SHA-256 body_sha256 revalidated on every read. TypeScript port of the primitive proven in erp-control-plane; canonical JSON via @cubiczan/chp.
Distribution Surfaces
- packages/npm/: npm-friendly manifest and catalog export, published as @cubiczan/trust-ledger-os@0.1.2.
- packages/python/: PyPI-friendly manifest and catalog export, published as trust-ledger-os==0.1.2.
- packages/mcp/: MCP-ready tool and resource catalog, published as trust-ledger-os-mcp@0.1.2.
Package Commands
- npm package build: npm --prefix packages/npm run build
- MCP package build: npm --prefix packages/mcp run build
- MCP server run: npm --prefix packages/mcp start
- Python CLI: PYTHONPATH=packages/python/src python3 -m trust_ledger_os.cli --manifest