conventions-mcp

Personal memory for coding conventions and standing instructions — one store, any MCP-compatible AI client, available in every project. Occasional other notes are fine too, but this is primarily meant to hold things like "always use 2-space indent in this language," "never force-push to main," corrections after getting something wrong, and workflow preferences — the kind of thing that should carry across sessions and projects rather than get re-explained every time.

Why this one

There's no shortage of memory MCP servers — several well-established ones (mem0/OpenMemory, Zep/Graphiti, the official reference memory server, plus a long tail of smaller projects) already do "remember things across sessions." What's different here:

• Narrow taxonomy, not a general note-taking store. Every capture gets forced into one of five purpose-built types — convention, instruction, correction, preference, other — plus a scope and a project field. A general "remember anything" store gives you a pile of loosely-related notes to search through; this one is opinionated about what belongs in it at all, which keeps retrieval precise instead of noisy.

• Deterministic retrieval, not best-effort. Most memory MCPs rely entirely on the calling model noticing a tool description is relevant and deciding to call it — which fails silently and inconsistently. Three Claude Code hooks here (SessionStart, UserPromptSubmit, PreToolUse) instruct a list_rules call on a fixed schedule — before the model's first action, and again every turn — and the last one actually enforces it by denying other tool calls until list_rules has been attempted that turn, rather than just repeating the request. If you're not on Claude Code, you still get the tool descriptions, just not the hook guarantee.

• Transparent by default, not silent. Every capture and update echoes the verbatim stored content and its scope back immediately, so a misheard or misclassified rule is visible and correctable on the spot — not something you discover three sessions later via search.

• Fully local, zero network calls. SQLite + local embeddings, no hosted service, no per-token costs, no API key. Classification (type/scope/topics/project) is done by the calling agent at capture time, guided by the tool description — it already has the full conversation the thought came from, richer context than an isolated content string would give a separate extractor model.

• Project-scoped without fuzzy matching. A rule can be global (the default) or tied to one specific codebase, and which project it belongs to is derived from the actual git repo, not guessed by an LLM from free text — so retrieval doesn't depend on an LLM having phrased a project name consistently across captures.

If what you want is a general-purpose "remember everything" store, or you're not on Claude Code and don't need the hook-driven determinism, one of the more general options above may fit better. This one is for someone who specifically wants a tight, coding-convention-focused memory that stays accurate and doesn't require trusting the model to remember to check it.

Related MCP server: myBrAIn

What it's tuned to store

Every capture is classified into one of five types by the calling agent, guided by capture_thought's tool description (src/server.js):

Each thought also gets a scope (a language/framework/topic, or "global") and 1–3 topic tags for filtering. This is deliberately narrow — it's not a general note-taking store — but the taxonomy isn't hardcoded logic, it's just the wording of the tool description and its zod schema in src/server.js. Retuning what counts as a convention vs. an instruction, adding a new type, or changing what "scope" means is a matter of editing that description text, not restructuring the code. The one wrinkle: the five type names are also referenced in the type filter's enum in list_thoughts (src/server.js) — if you rename or add a type, update that enum too or the new type will get rejected as a filter value. Everything's stored as a JSON blob column, so none of this needs a schema migration.

Separately, every thought gets a project field — null by default (applies everywhere), or a specific project name if it's obviously scoped to one codebase. The calling agent only judges whether it's project-specific; the actual project name is derived deterministically from the current git repo's directory name, not guessed by the model — so retrieval can do an exact match instead of fuzzy text comparison.

• Storage: SQLite (better-sqlite3) + sqlite-vec for native vector search, FTS5 for keyword search, combined via reciprocal rank fusion. One file, no server, no daemon.

• Embeddings: local, via Xenova/bge-small-en-v1.5 (384-dim, quantized, ~130MB). Loads lazily on first use, no network call, no GPU needed.

• Classification: done by the calling agent (Claude Code, or any MCP client) at capture time, guided by the tool description — no network call, no external model, no API key.

• Transport: MCP over stdio. No port, no listener, no CORS, no shared secret — the trust boundary is simply "who can launch this process," same as any other local tool.

• Proactive retrieval: two Claude Code hooks (see below) load and re-remind about standing rules deterministically every session and every turn — no CLAUDE.md instruction to keep in sync, no dependence on the model happening to notice a tool description is relevant.

Setup

Two ways to get this: a git checkout (if you want to read/modify the source) or the npm package (if you just want it running).

Git checkout:

npm install
npm run init-db        # creates data/memory.db

npm package:

npm install -g conventions-mcp
conventions-mcp init-db    # creates ~/.conventions-mcp/memory.db

Nothing to configure — there's no API key and no external service. MEMORY_DB_PATH is the only environment variable this reads, and it's optional (see .env.example).

Register with Claude Code

Register at user scope so it's available in every project, not just one repo — use the claude mcp add CLI, not a hand-edited config file:

# Git checkout — an absolute path, since Claude Code may spawn this from an
# arbitrary working directory:
claude mcp add --scope user conventions -- node /absolute/path/to/conventions-mcp/src/server.js

# npm package — already on PATH:
claude mcp add --scope user conventions -- conventions-mcp

Either way, this writes to ~/.claude.json's mcpServers key, which is what the CLI actually reads; a mcpServers entry placed directly in ~/.claude/settings.json is silently inert. Verify with claude mcp list. A new Claude Code session is required to pick up a newly-registered server.

Standing-rule hooks

Three hooks in ~/.claude/settings.json make retrieval deterministic every session and every turn — the first two instruct the model, the third actually enforces it:

{
  "hooks": {
    "SessionStart": [
      { "hooks": [{ "type": "command", "command": "node /absolute/path/to/conventions-mcp/bin/session-rules.js", "timeout": 15 }] }
    ],
    "UserPromptSubmit": [
      { "hooks": [{ "type": "command", "command": "node /absolute/path/to/conventions-mcp/bin/prompt-reminder.js", "timeout": 5 }] }
    ],
    "PreToolUse": [
      { "matcher": "*", "hooks": [{ "type": "command", "command": "node /absolute/path/to/conventions-mcp/bin/pre-tool-check.js", "timeout": 5 }] }
    ]
  }
}

• bin/session-rules.js instructs the model, before its first action each session, to call the list_rules tool — rather than embedding rule content in the hook output directly, which doesn't scale (a large enough stored rule set gets silently truncated to a small preview before it ever reaches the model).

• bin/prompt-reminder.js fires on every turn, re-issuing that same list_rules instruction fresh (so it can't scroll out of context the way text loaded once at session start could) and re-anchoring "capture this if it's a new rule" — there's no hook event for "the user just stated a rule," since that's a semantic judgment only the model can make.

• bin/pre-tool-check.js fires before every tool call and denies it outright unless list_rules has already been attempted this turn — the first two hooks are advisory (a model can ignore injected text), so this is the layer that actually closes the gap instead of just repeating the ask.

None of the three touch the database directly — list_rules itself resolves the current project from the session's working directory when the model calls it, so this works correctly regardless of where the conventions-mcp install itself lives on disk.

npm package: the scripts live inside the global install rather than a known clone path — resolve it first with npm root -g, then point the hook at $(npm root -g)/conventions-mcp/bin/session-rules.js the same way.

Windows: point the command at the .cmd wrapper instead of the .js file directly (no node prefix — the batch file invokes it) — bin\session-rules.cmd / bin\prompt-reminder.cmd / bin\pre-tool-check.cmd for a git checkout, or the equivalent path under npm root -g for the npm package.

Tools

Notes

• If a single message states several distinct rules, capture_thought gets called once per rule, each relayed individually — not merged into one capture or summarized together.

• The database lives at data/memory.db in a git checkout, or ~/.conventions-mcp/memory.db for the npm package (override either with MEMORY_DB_PATH). It's gitignored — back it up yourself if you care about it (it's just a SQLite file; cp is a complete backup).

• To upgrade embedding quality later without re-architecting, swap MODEL_NAME in src/embeddings.js — but re-embed existing thoughts if the new model's vector space isn't compatible with the old one (different models' embeddings aren't comparable, even at the same dimension).