Every AI memory framework stores and searches memories. But storing ≠ using. PMS solves the second problem — a memory consumption methodology for Claude Code and OpenClaw that transforms stored memories into actionable lessons, so your agent gets measurably better at helping you work. Built on 4 Hooks + 1 MCP server + 1 indexer.
Stars
2
Forks
0
Watchers
2
Open Issues
0
Overall repository health assessment
No package.json found
This might not be a Node.js project
41
commits
feat: add OpenClaw integration — Skill, Plugin (4 hooks), templates, and docs
12e6425View on GitHubrefactor: move role definition from CLAUDE.md to soul.md for model-agnostic design
599c51fView on GitHubdocs: add prompt engineering standards + restructure soul.md with XML
920d1d0View on GitHubfeat: add SOUL.md identity layer and agent persona onboarding
8e06250View on GitHubfeat: implement tiered retrieval for session summary → transcript tracing
53c6874View on GitHubfeat: add conversation history indexing, tiered retrieval, and rewrite README
3a2e43fView on GitHubfix: remove remaining "ego" references in customization docs
81b6b63View on GitHubfix: replace real product name in examples with generic placeholders
0f7bad1View on GitHubfix: improve bootstrap profile extraction quality for roles, projects, and placeholders
cc8fe2cView on GitHubfix: unify _dir → _path variable names for Path objects in auto-install.py
dd177b8View on GitHubUpdate README directory tree to show projects/ in all domains
0d3e699View on GitHubRemove obsolete bootstrap summary docs and clean up writer
24c0167View on GitHub