Obsidian Brain

Paste it into your assistant. It asks for the manifest, the server wiring and the skills, and for an honest account of what this record does not say. Read that second file first.

Draft an Agent Plugin (agent-plugins.org, specification 1.1.0) that wraps the
MCP server described below, so that somebody could install one thing and have
an assistant that knows when and how to use it.

An Agent Plugin is one installable unit: a `plugin.json` manifest, an
`mcp.json` that wires up the servers it needs, and a `skills/` directory
where each skill is a folder holding a `SKILL.md`. Hand back every file in
full, each under its own path, ready to save.

1. Write `plugin.json` with `$schema` exactly `https://agent-plugins.org/schemas/1.1.0/plugin.schema.json`. The name is
   1 to 64 characters of a-z, 0-9, `-` and `.`, alphanumeric at both ends, with
   no `--` and no `..` in it.

2. Write `mcp.json` wiring THIS server exactly as its record declares it. A
   remote keeps the URL and the transport type as written. A package keeps the
   registry, the identifier and the version as written. Do not invent a command,
   a port, a flag or an argument that is not in the record.

3. Every secret stays an input. No key, token, password or connection string
   belongs in either file. Declare what has to be supplied, name it, and say what
   it is for.

4. Do not invent tools. The record lists the tools it lists, and if it lists
   none then the honest plugin says the tool list was not published rather than
   guessing one from the description.

5. Skills are jobs, not tools. Write one skill per thing somebody would actually
   ask for, and inside each one say when to reach for this server, what a good
   result looks like, and what to do when it comes back empty. A skill per tool
   is a manual page with a different filename.

6. Say where it stops. Name what this plugin will not do — what it has no tool
   for, what needs a person, and what it must not be pointed at. A plugin with no
   stated edge reads as one with no edge.

7. Keep the author's own words for the description. If you would rather say it
   differently, say yours somewhere else and leave theirs where it is.

8. Record which version of the server you wrapped, and where the record came
   from, at the top of `plugin.json`'s description or in the readme. A plugin
   nobody can trace back to a version is one nobody can update.

Produce a second file alongside them, `LIMITS.md`, and treat it as the more
important of the two. The plugin is for whoever installs it. This is for
whoever has to decide whether installing it is a good idea, and that is
usually a different person who will never read the manifest.

It has three parts.

**What this is built from.** One paragraph: whose server it is, what the
record says it does, which version, and the fact that the record is all you
had. Say plainly that nobody ran it.

**What the record does not say.** One entry per gap. Whether the tool list was
published. What the server does with what it reads. What it costs. Whether it
writes anything anywhere. What credentials it will ask for and what those
credentials can reach. An unanswered question stays an unanswered question:
do not fill one in from the description or from what similar servers usually do.

**What a person has to check before trusting it.** The specific things
somebody should verify for themselves, in the order that would stop them
soonest if the answer is bad.

Write it in plain English, and do not soften it. A plugin drafted from a
directory record is a starting point to argue with, not something to install
into anything that matters.

The server record follows, exactly as the public index holds it. It is
everything I have: nobody has run this server, called a tool on it, or checked
that the address answers.

```json
{
  "server": {
    "$schema": "https://static.modelcontextprotocol.io/schemas/2025-12-11/server.schema.json",
    "name": "io.github.sweir1/obsidian-brain",
    "description": "Obsidian MCP server: semantic search, knowledge graph, and vault editing. No plugin required.",
    "title": "Obsidian Brain",
    "repository": {
      "url": "https://github.com/sweir1/obsidian-brain",
      "source": "github"
    },
    "version": "1.7.24",
    "websiteUrl": "https://sweir1.github.io/obsidian-brain/",
    "packages": [
      {
        "registryType": "npm",
        "identifier": "obsidian-brain",
        "version": "1.7.24",
        "runtimeHint": "npx",
        "transport": {
          "type": "stdio"
        },
        "runtimeArguments": [
          {
            "value": "-y",
            "type": "positional"
          }
        ],
        "packageArguments": [
          {
            "value": "server",
            "type": "positional"
          }
        ],
        "environmentVariables": [
          {
            "description": "Absolute path to your Obsidian vault (or any folder of .md files).",
            "isRequired": true,
            "format": "string",
            "name": "VAULT_PATH"
          },
          {
            "description": "Where to store the SQLite index + embedding cache. Defaults to $XDG_DATA_HOME/obsidian-brain or ~/.local/share/obsidian-brain.",
            "format": "string",
            "name": "DATA_DIR"
          },
          {
            "description": "Preset name: english (default, bge-small-en-v1.5), english-fast, english-quality, multilingual, multilingual-quality, multilingual-ollama. Ignored when EMBEDDING_MODEL is set.",
            "format": "string",
            "default": "english",
            "choices": [
              "english",
              "english-fast",
              "english-quality",
              "multilingual",
              "multilingual-quality",
              "multilingual-ollama"
            ],
            "name": "EMBEDDING_PRESET"
          },
          {
            "description": "Power-user override: any transformers.js checkpoint or Ollama model id. Takes precedence over EMBEDDING_PRESET. Switching auto-reindexes.",
            "format": "string",
            "name": "EMBEDDING_MODEL"
          },
          {
            "description": "Embedding backend. 'transformers' (local, default) or 'ollama' (requires a running Ollama server).",
            "format": "string",
            "default": "transformers",
            "choices": [
              "transformers",
              "ollama"
            ],
            "name": "EMBEDDING_PROVIDER"
          },
          {
            "description": "Base URL of a local Ollama server. Only used when EMBEDDING_PROVIDER=ollama.",
            "format": "string",
            "default": "http://localhost:11434",
            "name": "OLLAMA_BASE_URL"
          },
          {
            "description": "Override the embedding dimensionality when EMBEDDING_PROVIDER=ollama. If unset, the server probes the model on startup.",
            "format": "number",
            "name": "OLLAMA_EMBEDDING_DIM"
          },
          {
            "description": "Override Ollama's num_ctx for embed requests. Leave UNSET to let obsidian-brain auto-detect via `/api/show`'s `context_length` (e.g. nomic-embed-text=2048, bge-m3=8192, qwen3-embedding:0.6b=32 768). Setting this manually imposes a hard cap and may silently truncate longer inputs — see https://github.com/ollama/ollama/issues/14259. Ollama's own default is 2048 which truncates for any model trained on a larger context. See also https://github.com/ollama/ollama/issues/7008. Fallback when both env is unset AND /api/show is unreachable: 8192.",
            "name": "OLLAMA_NUM_CTX"
          },
          {
            "description": "Auto-pull the configured Ollama model when /api/show returns 404 (model not present). Default ON — choosing an Ollama-backed preset is implicit consent to download its model. Streams /api/pull progress to stderr. Set to '0' to disable auto-pull entirely (master kill-switch) and fall back to the actionable error path (`HTTP 404 — try: ollama pull <model>`).",
            "format": "string",
            "choices": [
              "",
              "0",
              "1"
            ],
            "name": "OBSIDIAN_BRAIN_OLLAMA_AUTO_PULL"
          },
          {
            "description": "Opt-in to auto-pull for BYOM (custom EMBEDDING_MODEL) Ollama models OUTSIDE Ollama's official `library/` namespace. Default OFF — third-party models (e.g. `user/custom-fork`, `myregistry.com/team/model`) require this env var to be set to `1` before they will auto-pull, to prevent silent downloads of arbitrary user-named artifacts. Preset-known models (via EMBEDDING_PRESET) and bare model ids in the official library (e.g. `qwen3-embedding:0.6b`, `library/llama3:8b`) continue to auto-pull by default per the existing OBSIDIAN_BRAIN_OLLAMA_AUTO_PULL behavior.",
            "format": "string",
            "choices": [
              "",
              "0",
              "1"
            ],
            "name": "OBSIDIAN_BRAIN_OLLAMA_BYOM_AUTO_PULL"
          },
          {
            "description": "Set to '1' to disable the live chokidar file watcher. Useful on SMB/NFS vaults where FSEvents/inotify don't fire reliably — fall back to running `obsidian-brain index` on a schedule (launchd/systemd).",
            "format": "string",
            "name": "OBSIDIAN_BRAIN_NO_WATCH"
          },
          {
            "description": "Set to '1' to skip the startup catchup reindex pass that picks up edits made while the server was down. The live file watcher still starts (via OBSIDIAN_BRAIN_NO_WATCH=1 to disable that separately), and first-time indexing on an empty DB is unaffected — this knob only governs the post-restart `enqueueBackgroundReindex` walk.",
            "format": "string",
            "name": "OBSIDIAN_BRAIN_NO_CATCHUP"
          },
          {
            "description": "Per-file reindex debounce for the live watcher, in milliseconds.",
            "format": "number",
            "default": "3000",
            "name": "OBSIDIAN_BRAIN_WATCH_DEBOUNCE_MS"
          },
          {
            "description": "Graph-wide community-detection (Louvain) debounce for the live watcher, in milliseconds. Louvain is the only expensive op — batching it prevents per-edit CPU spikes.",
            "format": "number",
            "default": "60000",
            "name": "OBSIDIAN_BRAIN_COMMUNITY_DEBOUNCE_MS"
          },
          {
            "description": "Per-tool-call timeout in milliseconds. Tools exceeding this return an MCP error instead of hanging.",
            "format": "number",
            "default": "30000",
            "name": "OBSIDIAN_BRAIN_TOOL_TIMEOUT_MS"
          },
          {
            "description": "Override the adaptive chunk-size budget (in tokens). When set, this beats the capacity probed from the model's tokenizer or Ollama /api/show. Use for debugging or for models with stale tokenizer configs.",
            "format": "number",
            "name": "OBSIDIAN_BRAIN_MAX_CHUNK_TOKENS"
          },
          {
            "description": "Override the per-user config directory where obsidian-brain stores model overrides (`model-overrides.json`) and the user-fetched seed (`seed-models.json`). Default is `$XDG_CONFIG_HOME/obsidian-brain` on macOS/Linux (or `~/.config/obsidian-brain`) and `%APPDATA%/obsidian-brain` on Windows. Both files survive `npm update obsidian-brain` because they live outside the package.",
            "format": "filepath",
            "name": "OBSIDIAN_BRAIN_CONFIG_DIR"
          },
          {
            "description": "Set to \"1\" to print a verbose synchronous startup trace to stderr — every preflight, createContext, server.connect, and shutdown step is logged with a monotonic timestamp. The LAST line before any silent failure tells you exactly which step the server reached. No-op when unset (no overhead). Use to diagnose silent-crash failure modes.",
            "format": "string",
            "choices": [
              "",
              "1"
            ],
            "name": "OBSIDIAN_BRAIN_DEBUG"
          },
          {
            "description": "Set to 'ndjson' for one-JSON-object-per-line stderr output (timestamp + level + message + structured fields). Default is human-readable plain text (`obsidian-brain: <message>`). Useful for piping logs into aggregators (Datadog, Loki, Vector, journald) that index structured fields.",
            "format": "string",
            "choices": [
              "",
              "ndjson"
            ],
            "name": "OBSIDIAN_BRAIN_LOG_FORMAT"
          }
        ]
      }
    ]
  },
  "_meta": {
    "io.modelcontextprotocol.registry/official": {
      "status": "active",
      "statusChangedAt": "2026-05-16T15:24:45.386473Z",
      "publishedAt": "2026-05-16T15:24:45.386473Z",
      "updatedAt": "2026-05-16T15:24:45.386473Z",
      "isLatest": true
    }
  }
}
```

The server is listed as "Obsidian Brain".