agentmem

stdionone★ 2development

Governed memory for coding agents with trust lifecycle, conflict detection, staleness tracking, and health scoring. SQLite + FTS5, zero infrastructure. Works with Claude Code, Cursor, Codex, Windsurf.

github ↗

Install

How to run this server

[ { "cmd": "pip install quilmem[mcp]", "imports": [] } ]

server path: agentmem

Tools

What this server exposes

memory_health

Run a health check on the memory system to confirm it's working

memory_add

Store a typed memory with status (hypothesis/active/validated/deprecated/superseded)

memory_search

Search memories by text, automatically ranking validated and active memories highest

memory_recall

Recall best memories fitting within a token budget, ranked by trust level

memory_promote

Promote a memory from hypothesis to active, or active to validated

memory_deprecate

Mark a memory as deprecated with a reason

memory_supersede

Replace an outdated memory with a newer one, linking old to replacement

detect_conflicts

Find memories that contradict each other using topic overlap and negation matching

detect_stale

Find outdated memories by age, missing source files, or hash drift

health_check

Score memory system health 0-100 based on conflicts, stale percentage, orphaned references, deprecated weight, and validated memories

Configuration

Environment & auth

authnone

Resources

githubhttps://github.com/Thezenmonster/agentmem ↗