skillmem

Self-improving skills for Claude Code and Codex — your agents learn, recall, reinforce, and forget.

skillmem gives Claude Code and the Codex CLI a local, persistent skill & memory layer. After every non-trivial task the agent can record how it was done as a skill; before the next task it recalls the relevant ones; skills that keep proving useful get stronger, and skills nobody uses fade away — the way human memory works.

• $0 per write and per read — no LLM calls, no cloud, no API keys. Plain SQLite on your disk.
• Bilingual hybrid search, fully local — FTS5 BM25 + Snowball stemming (EN/RU) + a multilingual ONNX embedding model. A Russian query finds an English skill and vice versa, all on CPU, offline.
• Ebbinghaus strength model — reinforce bumps a skill's strength, scheduled decay fades unused ones, lifecycle sweeps move dead skills to a backed-up archive (never deleted).
• Tamper-evident history — every edit is appended to a SHA256 hash-chain; skillmem verify detects any after-the-fact tampering.
• Deep Claude Code integration — 6 hooks + 8 MCP tools installed with one command.
• One memory, several agents — Claude Code and Codex share a single database, and every record carries the agent that wrote it, taken from the MCP handshake, so authorship stays readable when they learn side by side.
• Cross-platform — macOS (launchd), Windows (schtasks), Linux (systemd user timers, cron fallback).
• No vendor lock — export-all dumps everything to plain markdown with YAML frontmatter; re-importing the dump yields the same records.

Why

Agents repeat their mistakes because each session starts from zero. Existing "memory" tools store facts; skillmem stores procedures — trigger, steps, outcome, lessons — and ranks them by how often they actually helped. The write path costs nothing, so the agent can afford to learn from every task.

Quickstart

macOS / Linux:

bash install.sh                 # installs python + uv if needed, venv, symlinks

Windows (PowerShell):

powershell -ExecutionPolicy Bypass -File install.ps1

Or from a checkout:

uv venv && uv pip install -e '.[semantic]'
source .venv/bin/activate       # or prefix the commands below with `uv run`
skillmem init --claude-code     # wires MCP server + hooks into Claude Code
skillmem init --codex           # wires the MCP server into the Codex CLI
skillmem doctor                 # health check: DB, schema, semantic status

Both flags can be combined in one run — the two agents then share one database.

init --claude-code registers the MCP server in ~/.claude.json and the hooks in ~/.claude/settings.json (idempotent, with backups). Use --hooks minimal for just the Stop→migrate hook, or --hooks none for MCP only.

Codex CLI

skillmem init --codex

Appends an [mcp_servers.skillmem] table to ~/.codex/config.toml and marks the entry with SKILLMEM_AGENT=codex. The tag is belt-and-braces: with no tag set, the server takes the author's name from the agent's own MCP handshake, so attribution is right in a shared database whichever way skillmem was installed. The file is appended to, never rewritten: your own settings and comments stay where you put them, the result is parsed before it is written, and invalid TOML is refused rather than overwritten. skillmem uninstall removes the table again and leaves the rest of the file intact.

Codex reads AGENTS.md for project rules; if you keep yours in CLAUDE.md, point Codex at it with project_doc_fallback_filenames = ["CLAUDE.md"] in the same config file — then both agents follow one set of rules and one memory.

As a plugin

The repo is also a plugin, in two flavours, both pointing at the same skillmem-mcp binary:

• Agent Plugins (plugin.json + mcp.json at the repo root) — what the Codex CLI installs from a marketplace. mcp.json needs both its $schema and "type": "stdio", and the command must be a bare executable name rather than an absolute path — Codex's parser ignores the file otherwise, with no error. codex mcp list listing the server is the check that it parsed.
• Claude Code (.claude-plugin/ + hooks/hooks.json) — MCP server and all six hooks in one install.

Either way the package itself must be on PATH (pip install skillmem); the plugin wires the server, not the runtime. An MCP Registry manifest (server.json) is in the repo as well:

/plugin marketplace add liza-studio/skillmem
/plugin install skillmem@liza-studio

The plugin requires the skillmem Python package on PATH and replaces skillmem init --claude-code's wiring — use one or the other, not both (see docs/PUBLISHING.md).

Claude Desktop (chat app)

The MCP server also works in the Claude Desktop chat app — add to claude_desktop_config.json (Settings → Developer → Edit Config):

{
  "mcpServers": {
    "skillmem": { "command": "skillmem-mcp" }
  }
}

You get all 8 mem_* tools on demand (search, learn, recall, reinforce…). The automatic hooks (auto-recall on every prompt, session recap) are a Claude Code mechanism and do not run in the chat app.

How it works

 learn ──▶ recall ──▶ reinforce ──▶ decay
   │          │            │           │
   │          │            │           └─ daily job: unused skills lose strength;
   │          │            │              fully faded ones are archived (backed up)
   │          │            └─ strength +0.15 when a skill proves useful
   │          └─ hybrid BM25 + vector search, strength-weighted ranking
   └─ after a hard task: trigger / steps / outcome / lessons

1. learn — after a task that took real debugging, the agent calls mem_learn with a slug, trigger, steps, outcome, and lessons.
2. recall — before the next task, mem_recall (or the automatic hooks) surfaces the most relevant skills, fusing lexical and semantic signals via Reciprocal Rank Fusion.
3. reinforce — when a recalled skill helped, mem_reinforce bumps its strength, so proven skills rank higher next time.
4. decay — a scheduled skillmem decay run applies Ebbinghaus-style forgetting; skills untouched for months drift to stale, then to an archived state (excluded from recall, restorable with one command, snapshotted to JSONL first).

MCP tools

Hooks

All hooks are best-effort: a broken database or missing model never blocks Claude Code.

CLI highlights

skillmem learn skill-x -t "..." --trigger "..." --steps "..." --outcome success
skillmem recall "deploy the bot to prod"
skillmem skills                  # list skills with strength bars
skillmem decay --days 14         # manual decay + lifecycle sweep
skillmem search "hash chain" --kind feedback
skillmem verify --strict         # check the tamper-evidence chain
skillmem export-all ./vault      # markdown round-trip, no lock-in
skillmem import-vault ~/Obsidian/Notes
skillmem schedule install        # decay daily 04:15, export weekly Sun 04:30

Uninstall

skillmem uninstall               # removes MCP entries (both agents), hooks, scheduled jobs; keeps the DB
skillmem uninstall --purge-db    # ...and deletes the database

Config edits are made atomically with timestamped backups; corrupt JSON or TOML is never overwritten.

Benchmarks

Retrieval quality on LongMemEval (Wu et al., ICLR 2025), full oracle set, hybrid retrieval (FTS5 BM25 + Snowball stemming + paraphrase-multilingual-MiniLM-L12-v2 embeddings, RRF fusion), k=5, CPU only:

Median 0.76 s per query on a laptop CPU, no LLM calls, no network. The pipeline is deterministic: repeated runs produce identical numbers. Reproduce with python bench/longmemeval.py --sample 0 -k 5 (see bench/README.md for the oracle file and reporting rules — we don't publish bare percentages without stating the retrieval mode and embedding model, and we encourage other tools to do the same).

License

Apache-2.0 — see LICENSE.

Built by Liza Studio.