| name | mcp-scripting |
| description | Write mcp_script JavaScript for discovering, inspecting, and calling MCP tools. |
MCP scripting
For multi-call MCP work, write ordinary JavaScript with loops, filtering, chaining, fan-out, or other logic between calls. Run that source with mcp_script; it is the primary MCP orchestration surface. For a single MCP search, describe, status check, auth action, or tool call, use mcp instead.
Write the source naturally, then pass it as mcp_script's code argument:
const { items } = await tools.search({ query: "search issues", server: "github" });
const candidate = items[0];
if (!candidate) return { error: "No matching tool" };
const details = await tools.describe({ path: candidate.path });
if (details.error) return details;
const result = await tools.call(details.path, { query: "is:open label:bug" });
if (!result.ok) return result;
emit({ tool: details.path, completed: true });
return result.data?.structuredContent ?? result.data;
Workflow
- Find candidate tools with
await tools.search({ query, server?, limit?, offset? }). limit defaults to 12 and is capped at 100.
- Inspect the exact returned path with
await tools.describe({ path }).
- Call it with
tools.call(path, args).
Calls resolve to { ok: true, data } or { ok: false, error }; handle failed calls instead of expecting them to stop the script. Successful data is the raw MCP result: tool calls usually return { content, structuredContent? }, while resource reads return { contents }. Prefer structuredContent when present, and only parse a text block when that server documents JSON text. emit(value) adds user-visible output before the final return value. console output is captured too.
tools is a non-enumerable proxy: Object.keys(tools) throws. Always use tools.search for discovery. When a known flat path is a valid identifier, direct calls such as tools.github_search_issues(args) are supported; use bracket syntax for hyphenated names: tools["server_tool-name"](args). search, call, describe, and promise/serialization names (then, catch, finally, toJSON, toString, valueOf) are reserved on the proxy; if a flat path collides with one, call it via tools.call("exact-path", args).
tools.search and tools.describe are asynchronous and must be awaited. The default script timeout is 30 seconds; the worker is terminated at the deadline, including for infinite loops. Every invocation still uses normal lazy connection, authentication, output guarding, and approval gates. Raw call results are cloned into the worker so scripts can reduce them before the model-facing guard; request scoped data instead of unnecessarily large payloads. Result details contain a concise calls trace with every search, describe, and call operation; each entry includes its query or path, outcome, and duration.
Use plain JavaScript loops and Promise utilities for composition. The sandbox has no Node, filesystem, or network globals such as process, Buffer, or fetch. Fluent helpers such as tools.find(...).one(), tools.parallel(...), and tools.retry(...) are not provided.