Skip to main content
Glama
yuchi-chang

obsidian-mcp

by yuchi-chang

Evaluate JavaScript in Obsidian

obsidian_eval
Destructive

Execute JavaScript in Obsidian with full access to the app object. Use only when other tools cannot meet the need.

Instructions

Runs arbitrary JavaScript inside the running Obsidian instance with access to the app object. DANGEROUS: can read/modify any vault data and execute side effects. Use only when no narrower tool fits.

Input Schema

TableJSON Schema
NameRequiredDescriptionDefault
vaultNoVault name to target. Optional — defaults to the most recently focused vault.
codeYesJavaScript code to evaluate.
confirmNoSet to true to skip the interactive confirmation prompt. Use only when the caller has already confirmed with the user.

Implementation Reference

  • src/tools.ts:272-272 (registration)
    The 'tools' array is exported, and obsidian_eval is registered as part of this array (line 755). The tool definitions array is iterated over in src/index.ts (line 40-51) where each tool is registered with the MCP server.
    export const tools: ToolDef[] = [
  • Input schema for obsidian_eval: accepts 'vault' (optional, string), 'code' (required string of JavaScript to evaluate), and 'confirm' (optional boolean from ConfirmArg).
    inputSchema: {
      ...VaultArg,
      code: z.string().min(1).describe("JavaScript code to evaluate."),
      ...ConfirmArg,
    },
  • Handler function for obsidian_eval. Checks if the code exceeds the platform's argv limit using shouldChunk/getLimit. If too large, returns an error suggesting the user write the script to a note first. Otherwise, runs the code via runText with command 'eval'.
      handler: async ({ vault, code }) => {
        if (shouldChunk(code)) {
          return errorResult(
            new Error(
              `Eval code is ${Buffer.byteLength(code, "utf8")} bytes, over the ` +
                `${getLimit()}-byte argv limit on this platform. JavaScript cannot ` +
                `be chunked — write the script to a note via append (which auto-chunks) ` +
                `then load it inside a smaller eval, e.g. ` +
                `\`new Function(await app.vault.adapter.read('script.md'))()\`.`,
            ),
          );
        }
        return runText("eval", { vault, params: { code } });
      },
    },
  • src/index.ts:40-51 (registration)
    MCP server registration loop that registers each tool (including obsidian_eval) from the tools array using server.registerTool().
    for (const tool of tools) {
      server.registerTool(
        tool.name,
        {
          title: tool.title,
          description: tool.description,
          inputSchema: tool.inputSchema,
          annotations: tool.annotations,
        },
        buildHandler(server, tool),
      );
    }
  • runText helper function used by the obsidian_eval handler. Calls runObsidian with the given command and options, then formats the output as a text result.
    async function runText(
      command: string,
      opts: Parameters<typeof runObsidian>[1] = {},
    ): Promise<McpToolResult> {
      try {
        const result = await runObsidian(command, opts);
        const text = result.stdout.trim() || result.stderr.trim() || "(no output)";
        return textResult(text);
      } catch (err) {
        return errorResult(err);
      }
    }
Behavior4/5

Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?

The description warns that the tool 'can read/modify any vault data and execute side effects,' which aligns with annotations (destructiveHint=true, openWorldHint=true). It adds context about the `app` object beyond annotations.

Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.

Conciseness5/5

Is the description appropriately sized, front-loaded, and free of redundancy?

The description is extremely concise with two sentences: one for purpose and one for usage guidance. It is front-loaded with the core action and achieves maximal information density.

Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.

Completeness4/5

Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?

Given the tool's complexity and lack of output schema, the description covers purpose, danger, and usage guideline. It could mention return value format, but it remains fairly complete for a powerful tool.

Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.

Parameters3/5

Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?

The input schema has 100% coverage for all three parameters. The description does not add significant detail beyond what the schema provides, so a baseline score of 3 is appropriate.

Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.

Purpose5/5

Does the description clearly state what the tool does and how it differs from similar tools?

The description clearly states the tool evaluates arbitrary JavaScript in Obsidian with access to the `app` object. It distinguishes itself from siblings by the verb 'runs' and the unique capability of code execution.

Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.

Usage Guidelines4/5

Does the description explain when to use this tool, when not to, or what alternatives exist?

The description explicitly says 'Use only when no narrower tool fits,' providing clear guidance on when to use this tool versus alternatives. It does not list specific sibling tools, but the context implies caution.

Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.

Install Server

Other Tools

Latest Blog Posts

MCP directory API

We provide all the information about MCP servers via our MCP API.

curl -X GET 'https://glama.ai/api/mcp/v1/servers/yuchi-chang/obsidian-mcp'

If you have feedback or need assistance with the MCP directory API, please join our Discord server