Skillget

samemind

mcp
Guvenlik Denetimi
Basarisiz
Health Gecti
  • License — License: MIT
  • Description — Repository has a description
  • Active repo — Last push 0 days ago
  • Community trust — 12 GitHub stars
Code Basarisiz
  • eval() — Dynamic code execution via eval() in bench/longmemeval/adapter/samemind_adapter.py
  • eval() — Dynamic code execution via eval() in bench/longmemeval/bridge/eval-server.mjs
  • process.env — Environment variable access in bench/longmemeval/bridge/eval-server.mjs
  • network request — Outbound network request in bench/longmemeval/bridge/eval-server.mjs
  • spawnSync — Synchronous process spawning in bin/samemind.mjs
  • process.env — Environment variable access in bin/samemind.mjs
Permissions Gecti
  • Permissions — No dangerous permissions requested

Bu listing icin henuz AI raporu yok.

SUMMARY

Your personal universal memory for every AI agent. Switch engines. Same mind. Git-native, zero-infra, plain markdown + MCP.

README.md

samemind

ci

Your personal universal memory for every AI agent. Switch engines. Same mind.

Git-native markdown bundle (no daemon, no required DB): identity, search, handoff, an append-only work ledger, and a kanban board in one place. Wire-compatible with Google OKF v0.1. Optional embeddings; BM25 always works offline.

Why not “just markdown + BM25”?

Typical git-markdown memory samemind
Wire format ad hoc OKF v0.1 export/import
Identity flat notes Identity / User / EngineRule → budgeted brief
Work separate tools event ledger + board in the same bundle
Engines often one client samemind install → 12 engines (adapters)
Capture — samemind capture (read-only session → inbox)

First use

npx samemind setup

Detects the agent, scaffolds a bundle if needed, wires the memory protocol into its instruction file, registers MCP, probes local embeddings (or stays on honest BM25). Interactive by default; --yes / --dry-run / --target <dir> available.

npx samemind init --demo          # fresh dir only
npx samemind recall "context budget"
npx samemind board
npx samemind brief --engine claude-code

Agent self-install protocol: INSTALL_FOR_AGENTS.md.

Global personal bundle (project + global recall, project wins on id collision):

npx samemind setup --global

Details: docs/full-guide.md § Global mode (archived long homepage).

Proof (commands agents actually run)

Command Job
setup / install Wire engines + MCP
recall / gde Search (BM25 / optional semantic)
brief / handoff Identity + work-state across sessions
board Kanban over tasks/plans/ideas
capture Pull engine transcripts → inbox/
ledger Append-only work events
serve MCP: memory_search, memory_get, memory_write_inbox, …
forget / export / import Hygiene + OKF packs

Full table and env vars: docs/full-guide.md § Tools.

MCP (stdio):

npx samemind serve
claude mcp add samemind -- npx samemind serve

Security perimeter (secret visibility, inbox-only writes, path safety): docs/full-guide.md § MCP.

Docs map

Topic Doc
Full previous homepage (deep dive) docs/full-guide.md
Engine matrix / OpenClaw·Hermes bootstrap docs/adapters.md
Memory protocol (recall → cite → inbox) docs/memory-protocol.md
Identity + brief docs/identity-layer.md
Hygiene, supersedes, heat docs/memory-hygiene.md
Event ledger docs/event-ledger.md
Session capture docs/session-capture.md
Compaction / handoff docs/compaction-recipe.md
OKF interop docs/interop.md
Benchmark notes docs/benchmark.md

Limits (honest)

  • Canon promotion is human-gated (inbox → concepts); not auto-mem that rewrites truth silently.
  • Semantic search needs a local/OpenAI-compatible embeddings endpoint (OKF_EMBED_URL); without it, BM25 only — by design.
  • Hand-curated scale (roughly 10²–10³ concepts), not a 24/7 life-ingestion daemon — see vs gbrain.

Tests

node --test tools/*.test.mjs

License

MIT © 2026 Aleksandr Grebeshok

Yorumlar (0)

Sonuc bulunamadi