wikimem
llms.txt8 starsChanged 25 days agoMITView on GitHub
Self-improving memory for agents: ingest anything, query everything. A local-first Markdown wiki with a command line (wikimem), an MCP server (wikimem-mcp, 19 tools), three skills and a Claude Code plugin. Its agent employee is Codex. WikiMem turns files, links and connected apps into linked wiki pages, answers questions from them with citations, and scores and improves its own pages with an observer loop. Every vault is a folder of Markdown files, a git repository and an Obsidian vault.
In this repository31 other files it gives its agents.
- AGENTS.mdAGENTS.md
- CLAUDE.mdCLAUDE.md
- agent-runtimeSkill
- frontend-designSkill
- architectSkill
- deep-thinkSkill
- design-systemsSkill
- ecosystem-referenceSkill
- fractal-delegationSkill
- harness-reviewSkill
- heartbeatSkill
- integrate-resourcesSkill
- llm-council-critiqueSkill
- loop-createSkill
- mcp-setupSkill
- memory-compressionSkill
- memorySkill
- model-routingSkill
- oss-project-scaffoldSkill
- playwright-cliSkill
- repo-scoutSkill
- self-improveSkill
- skill-creatorSkill
- skill-routingSkill
- sprintSkill
- test-visualSkill
- troubleshootSkill
- ux-auditSkill
- wikimem-ingest-pipelineSkill
- wikimem-observerSkill
- wikimemSkill
The file
Files
# WikiMem > Self-improving memory for agents: ingest anything, query everything. A local-first Markdown wiki with a command line (`wikimem`), an MCP server (`wikimem-mcp`, 19 tools), three skills and a Claude Code plugin. Its agent employee is Codex. WikiMem turns files, links and connected apps into linked wiki pages, answers questions from them with citations, and scores and improves its own pages with an observer loop. Every vault is a folder of Markdown files, a git repository and an Obsidian vault. ## Start here - [AGENTS.md](AGENTS.md): what it does, every tool and how to call it, what not to do - [README.md](README.md): install from any coding agent, quick start, how it works - [SOUL.md](identity/SOUL.md): Codex, the agent employee behind WikiMem: persona, boundaries, operating model ## Skills - [wikimem](skills/wikimem/SKILL.md): ingest, search, ask, sync a connector, or expose a vault over MCP - [wikimem-ingest-pipeline](skills/wikimem-ingest-pipeline/SKILL.md): turn files, folders or URLs into linted, linked pages, incrementally - [wikimem-observer](skills/wikimem-observer/SKILL.md): score every page, find weak and missing pages, improve a budget-capped set ## MCP tools - [MCP server over stdio](src/mcp-server.ts): wikimem_search, wikimem_read, wikimem_list, wikimem_status, wikimem_ask, wikimem_ingest, wikimem_ingest_url, wikimem_pipeline, wikimem_scrape, wikimem_observe, wikimem_run_observer, wikimem_get_report, wikimem_improve, wikimem_lint, wikimem_connectors, wikimem_list_connectors, wikimem_connect, wikimem_sync, wikimem_preview - [MCP over HTTP with OAuth 2.1](src/mcp/http-server.ts): the same tools at `/mcp` when `wikimem serve` runs - [Claude Code plugin](.claude-plugin/plugin.json): skills, sub-agents, a vault hook and the MCP server in one install - [A2A agent card](.well-known/agent-card.json): name, description and skills for agent-to-agent discovery ## Docs - [Architecture](docs/architecture.md): the raw and wiki layers and the three automations - [Configuration](docs/configuration.md): `config.yaml`, providers, sources and schedules - [Connector architecture](docs/connector-architecture-reference.md): how connectors sync into a vault - [OAuth app registration](docs/OAUTH-APP-REGISTRATION.md): registering your own OAuth apps for connectors - [Evals](eval/README.md): the golden tasks (`pnpm eval`) and the weekly self-review - [Self-improvement contract](dsip.json): the loop, its evals and how a change is rolled back ## Optional - [Heartbeat](identity/HEARTBEAT.md): what runs on a schedule and what happens when it fails - [Memory](memory/MEMORY.md): long-term facts; [LEARNINGS.md](memory/LEARNINGS.md) holds corrections that became rules - [CHANGELOG.md](CHANGELOG.md): release history - [CONTRIBUTING.md](CONTRIBUTING.md) and [SECURITY.md](SECURITY.md)
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.