Download Latest Version
v3.1.2 source code.tar.gz (2.3 MB)
Get an email when there's a new version of ledgermind
| Name | Modified | Size | Downloads / Week |
|---|---|---|---|
| Parent folder | |||
| README.md | 2026-02-16 | 2.4 kB | |
| v1.11.0_ Granular Capabilities _ Knowledge Governance Hardening source code.tar.gz | 2026-02-16 | 64.0 kB | |
| v1.11.0_ Granular Capabilities _ Knowledge Governance Hardening source code.zip | 2026-02-16 | 103.3 kB | |
| Totals: 3 Items | 169.7 kB | 0 | |
GitHub Release Information
Release Title:
v1.11.0: Granular Capabilities & Knowledge Governance Hardening
Release Notes (Copy and Paste into GitHub):
Overview
This release introduces a major shift in the security model of the Agent Memory System, moving from simple role-based access to a granular Capability-based Access Control. It also includes significant stability improvements for the core logic and full documentation coverage in English.
📦 Component Versions
- agent-memory-server: v1.11.0
- agent-memory-core: v1.22.0
🚀 New Features & Enhancements
- Granular Capabilities: Replaced legacy RBAC with a flexible permission system. You can now define specific access for each agent:
read,propose,supersede,accept, andsync. - Secret-less Operation: Introduced the
--capabilities(JSON) and--capabilityCLI flags. This allows the server to run in trusted environments without requiring theAGENT_MEMORY_SECRETenvironment variable. - MCP Native Configuration: Added formalized configuration examples for connecting to MCP clients like Claude Desktop and VS Code.
🛡 Core Stability & Governance
- Strict Validation: Synchronized all components with Pydantic V2. Enforced minimum length requirements for critical fields (
rationaleandtarget) to prevent "low-effort" or vague agent entries. - Trust Boundary Hardening: Improved exception handling for
HUMAN_ONLYtrust boundaries, ensuring unauthorized write attempts are caught with explicitInvariantViolationerrors. - Git Audit Integrity: Refined the hybrid storage logic to ensure SQLite metadata and Git content stay perfectly synchronized even under high concurrency.
🧪 Quality Assurance
- Full Test Suite Passing: Fixed several legacy test failures caused by updated validation schemas.
- Core: 34/34 tests passing.
- Server: 10/10 tests passing.
📚 Documentation
- Full English Translation: All README files across the repository (root, core, and server) have been translated and updated to match the latest technical specifications.
- Quick-Start Guides: Updated installation and usage guides to reflect the new capability-based workflow.
Full Changelog: https://github.com/sl4m3/agent-memory/compare/v1.10.0...v1.11.0
