Caesar AI Agent Ledger (caesar-ai-agent-ledger)

Client-side SDK and runtime logging system for autonomous AI agent actions, part of the Caesar AI Governance Hub ecosystem.

---

📖 Overview

caesar-ai-agent-ledger is a runtime logging library and client-side SDK. It serves as a cryptographically signed execution ledger that tracks autonomous AI agent actions, prompt templates, tool executions, API endpoints, risk categories, and human-in-the-loop approvals.

This SDK is part of the Caesar AI Governance Hub ecosystem at caesar.no, providing a verifiable timeline of agent behavior.

🚦 Project Status

Note

This repository is currently in the runtime scaffold / local Supabase rehearsal + disabled local DB compile harness stub gate stage. Local validation, the metadata-first event model, local JSONL event buffer, local export/import dry-run contract, read-only local event projection, localhost-only read-only API, static private local dashboard, machine-checkable runtime boundary validation, local static Supabase migration rehearsal validation, and a disabled-by-default local DB compile harness stub are in place; persistence and any live event writing remain future gated work.

---

👥 Who It Is For

• AI Engineers & Framework Developers: Integrating continuous audit logging into agents built with LangChain, LlamaIndex, CrewAI, AutoGen, or custom runtimes.
• CTOs & Product Owners: Maintaining a reliable, tamper-resistant trail of autonomous decisions, database mutations, and transactions.
• Compliance & Risk Officers: Verifying that AI agents respect operating parameters and documented human-in-the-loop approvals.

---

🛠️ How It Connects

1. Caesar AI Governance Hub Connection

caesar-ai-agent-ledger is the runtime auditing arm of the ecosystem. It records the operational choices made by active agents, validating that real-time runtime flows respect the organizational policies monitored by the parent hub.

2. Connection to caesar-ai-evidence

All runtime logs and transaction records produced by the SDK conform strictly to the agent-run and related schemas defined in caesar-ai-evidence.

---

⚖️ Important Disclaimer

Important

No Compliance Guarantees: caesar-ai-agent-ledger is a logging utility and SDK designed to capture, sign, and document AI agent execution trails. It does not guarantee regulatory compliance, legal clearance, or audit approvals. Regulatory compliance remains a holistic legal, operational, and organizational state determined by accredited auditors, legal experts, and competent authorities.

---

📂 Repository Directory

• SPEC.md — SDK configuration guidelines, JSON log formats, and API scopes.
• ARCHITECTURE.md — Core SDK layout, cryptographic state signing, and execution flow.
• ROADMAP.md — Multi-phase project development roadmap.
• CHANGELOG.md — Chronological release history.
• REPO_INVENTORY.md — Structural file index of this codebase.
• PROJECT_STATE.md — Project phase, metadata tracker, and boundaries.
• NEXT_ACTIONS.md — Task execution lists and autonomous boundaries.
• docs/runtime/AGENT_EVENT_MODEL.md — Metadata-first event contract, IDs, risk levels, and retention rules.
• docs/runtime/LOCAL_EVENT_BUFFER.md — Local-only JSONL buffer behavior, validation, and test guidance.
• docs/runtime/LOCAL_EVENT_EXPORT_IMPORT_CONTRACT.md — Local-only bundle contract and import dry-run mapping rules.
• docs/runtime/READ_ONLY_LOCAL_EVENT_PROJECTION.md — Read-only projection contract for local buffer/exported events.
• docs/runtime/LOCALHOST_READ_ONLY_API.md — Localhost-only read-only API contract and safety boundaries.
• docs/runtime/STATIC_LOCAL_DASHBOARD.md — Static/private local dashboard contract and safety boundaries.
• docs/runtime/BACKEND_RUNTIME_READINESS_BOUNDARY.md — Backend runtime readiness boundary, invariants, forbidden actions, and approval gates.
• docs/runtime/LOCAL_SUPABASE_MIGRATION_REHEARSAL.md — Local-only static Supabase migration rehearsal boundary and compile-gate prerequisites (T022).
• docs/runtime/LOCAL_DB_COMPILE_HARNESS.md — Disabled-by-default local DB compile harness contract, stub behavior, and future approval prerequisites (T023).
• docs/RESEARCH_CONTEXT.md — Functional domain research and strategic context.
• docs/DECISION_LOG.md — Architectural decision log history.
• src/event-buffer/ — Local-only validator and JSONL buffer implementation.
• src/export-import/ — Local-only export bundle and import dry-run planner implementation.
• src/projection/ — Local-only read-only projection implementation over validated events.
• src/local-api/ — Localhost-only read-only HTTP API over the projection output.
• site/local-dashboard/ — Static private local dashboard UI and deterministic sample projection data.