OKF Agent Memory launches a Git-native Markdown memory layer for AI coding agents to preserve project context

OKF Agent Memory is an open source system for giving AI coding agents persistent memory inside a project repository. It stores architectural decisions, domain knowledge, and operational facts as plain Markdown files with YAML frontmatter under a knowledge/ directory, with Go-based tooling for parsing, validation, search, and MCP workflows.

The project reflects a broader shift toward repo-native AI infrastructure that keeps agent context auditable, portable, and version-controlled rather than locked in proprietary vector stores. If adopted, formats like OKF could make AI coding agents more reliable across sessions and vendors by standardizing how project knowledge is written, searched, and trusted.