MCP App Store
by KeenEthics
Script.it icon

Script.it MCP Server for ChatGPT

By Script.it

Overview

Give the work you do with ChatGPT a life beyond the conversation. Describe an automation, and your agent builds it as a script: a workflow of clear, reusable steps. Understand what each step does in everyday language, change its inputs without code, and run the whole workflow or just the step you need. Each step produces real files you can open, check, and share—from spreadsheets and PDFs to reports and images. Connect your business. Set your rules. Work across your everyday apps, internal tools, and even your legacy ERP. Start automations on a schedule, through a webhook, or when a specific event happens in your connected systems. Keep working while your laptop sleeps. Your agent has its own computer in the cloud, with files, a browser, and tools. Your automations keep running when you close ChatGPT or step away. Talk to your automations where you already work. Message your agent through email, Slack, or Telegram and get results back in the same thread. When a workflow needs your input, it can reach you there. Automate together, as one team. Share workflows, skills, and app connections in a common workspace. Teammates can run and improve the same automations without passing passwords around. Admins control which apps and actions each person—and their agents and automations—can use. Bring your favorite agent. Switch anytime. Use ChatGPT, Codex, Claude Code, or other compatible agents with the same Script.it workspace. Scripts use the open Agent Skills format, so you can read, share, and extend them with new steps, code, or skills.

How to connect

  1. ChatGPT: Open the Script.it listing in the ChatGPT app directory and choose Connect. ChatGPT listing Script.it exposes 17 tools here.

Tools

connect_integration

ChatGPT
Command shortcut: scriptit integrations connect <slug>. Read 'scriptit skills show script-authoring' through run_command for the full workflow, examples and command options. Get the Script.it connection-screen link for a catalog integration, including a custom integration type. Return the link to the user so they can sign in or enter credentials securely in Script.it. This operation does not connect the account itself and must not collect passwords, tokens or API keys in chat. After the user connects, list_integrations shows the usable connection. For general work or additional options, use the documented commands through run_command. Requires an existing Script.it session_id. Runs as a tracked command in that session. Read output and exit_code; a running command is not a completed operation. Continue pending or paginated output with read_output using the returned session_id, command_id and next_offset; do not repeat the operation to poll it.

connect_terminal

ChatGPT
Sign your own terminal in to the user's Script.it account for 24 hours, to drive Script.it with the scriptit CLI. Use this only when you need to move files between your machine and the sandbox and your terminal can reach the internet; otherwise use start_session and run_command. 1. Run scriptit auth status. If the command is missing, install the CLI with pip install scriptit-cli (the released package) or pip install git+https://github.com/script-it/cli.git (the latest main). If it shows a signed-in account, go to step 3. 2. Run scriptit auth login --no-browser --timeout 15 --api-url https://api.script.it. It prints a link and a code such as AB12-CD34, then exits saying the login is not approved yet; that is expected. Call this tool with that link and that code, exactly as printed. If the user let you sign terminals in, the tool approves the terminal: run the same login command again to finish signing in. Otherwise it answers with a link for the user to approve the terminal in Script.it; do what its answer says. 3. Run scriptit session new, then scriptit context. Read all of the context (write it to a file if your terminal truncates it) and carry out the request with the CLI as it describes. Never send the link the CLI printed anywhere but this tool; show the user only the link this tool returns. Never ask the user for a password or token.

create_custom_integration

ChatGPT
Command shortcut: scriptit integrations catalog create <slug> --auth-mode <mode> --base-url <url>. Read 'scriptit skills show custom-integrations' through run_command for the full workflow, examples and command options. Register an integration type for an API that is not already in the catalog. Supports API-key, bearer-token or basic authentication, an API base URL, documentation and an optional permission-policy file. The custom-integrations skill covers full definitions for OAuth, custom headers and other advanced cases through the documented API command. Creates a type, not a connected account; it does not accept credentials. By default the type is private; workspace shares the type with that workspace. Then use connect_integration to give the user the app's secure connection screen. For general work or additional options, use the documented commands through run_command. Requires an existing Script.it session_id. Runs as a tracked command in that session. Read output and exit_code; a running command is not a completed operation. Continue pending or paginated output with read_output using the returned session_id, command_id and next_offset; do not repeat the operation to poll it.

create_trigger

ChatGPT
Command shortcut: scriptit trigger create <catalog-id> --action '<run_script JSON>' --inputs '<settings JSON>'. Read 'scriptit skills show triggers' through run_command for the full workflow, examples and command options. Configure a saved script to respond to a schedule or an event, such as an incoming email, webhook or supported app change. Binds a trigger type from get_trigger_catalog to a stable script_id, the script's SKILL.md metadata.id (scriptit script <path> shows it). Supply its settings and required app connection. For schedule, settings include cron and time_zone, for example {"cron":"0 9 1","time_zone":"Asia/Jerusalem"}. Read 'scriptit skills show triggers --file SCHEDULED.md' for schedules, or '--file CREATING.md' for other events. Creates the trigger disabled and records its declaration in the script; it does not start future runs. After checking the workflow and configuration, use enable_trigger to activate the automation the user requested. For general work or additional options, use the documented commands through run_command. Requires an existing Script.it session_id. Runs as a tracked command in that session. Read output and exit_code; a running command is not a completed operation. Continue pending or paginated output with read_output using the returned session_id, command_id and next_offset; do not repeat the operation to poll it.

disable_trigger

ChatGPT
Command shortcut: scriptit trigger disable <id> --json. Read 'scriptit skills show triggers --file MANAGE.md' through run_command for the full workflow, examples and command options. Disable a schedule or event trigger to stop future dispatch while retaining its configuration and cursor. Does not cancel a workflow run that has already started. For general work or additional options, use the documented commands through run_command. Requires an existing Script.it session_id. Runs as a tracked command in that session. Read output and exit_code; a running command is not a completed operation. Continue pending or paginated output with read_output using the returned session_id, command_id and next_offset; do not repeat the operation to poll it.

enable_trigger

ChatGPT
Command shortcut: scriptit trigger enable <id-or-path> --json. Read 'scriptit skills show triggers --file MANAGE.md' through run_command for the full workflow, examples and command options. Enable a configured schedule or event trigger so future events can run its automation. Use only when the user requested activation and the workflow and required connections are ready. Pass the trigger_id, as disable_trigger takes it, or a script's declared trigger_path, which can create the user's own trigger row from that declaration. Enabling may resume event delivery and cause the script's external actions. Verify the returned state with list_triggers. For general work or additional options, use the documented commands through run_command. Requires an existing Script.it session_id. Runs as a tracked command in that session. Read output and exit_code; a running command is not a completed operation. Continue pending or paginated output with read_output using the returned session_id, command_id and next_offset; do not repeat the operation to poll it.

get_integration_catalog

ChatGPT
Command shortcut: scriptit integrations catalog list --search <text>. Read 'scriptit skills show custom-integrations' through run_command for the full workflow, examples and command options. Search available integration types before creating a new one or connecting an account. Returns platform and custom types, authentication mode and a next_cursor for more results. Lists the types the user can still connect: the included web search, AI model, social-media and Agent Email connections, and apps already connected with their one account, appear in list_integrations instead. For general work or additional options, use the documented commands through run_command. Requires an existing Script.it session_id. Runs as a tracked command in that session. Read output and exit_code; a running command is not a completed operation. Continue pending or paginated output with read_output using the returned session_id, command_id and next_offset; do not repeat the operation to poll it.

get_run_status

ChatGPT
Command shortcut: scriptit status <run-id> --session <session-id>. Read 'scriptit skills show scriptit-runtime' through run_command for the full workflow, examples and command options. Check a workflow's status, results, generated files, errors or pending questions. Pass the session_id containing the run, including the child session_id returned by run_script. Also reconciles a stale running record to failed if its process is gone. For general work or additional options, use the documented commands through run_command. Requires an existing Script.it session_id. Runs as a tracked command in that session. Read output and exit_code; a running command is not a completed operation. Continue pending or paginated output with read_output using the returned session_id, command_id and next_offset; do not repeat the operation to poll it.

get_trigger_catalog

ChatGPT
Command shortcut: scriptit trigger catalog --nojson; scriptit trigger catalog <id>. Read 'scriptit skills show triggers' through run_command for the full workflow, examples and command options. Discover supported ways to start an automation: schedules, incoming email, webhooks and events in connected apps. Lists a compact catalog of IDs, names, kinds and integrations. Pass a builtin_id for that type's settings schema and details. Use the returned catalog ID and settings schema when configuring create_trigger; do not invent event names. For general work or additional options, use the documented commands through run_command. Requires an existing Script.it session_id. Runs as a tracked command in that session. Read output and exit_code; a running command is not a completed operation. Continue pending or paginated output with read_output using the returned session_id, command_id and next_offset; do not repeat the operation to poll it.

list_integrations

ChatGPT
Command shortcut: scriptit integrations list --brief. Read 'scriptit skills show workspaces' through run_command for the full workflow, examples and command options. Find the user's usable app connections before building an automation across email, Slack, spreadsheets, CRMs or other services. Includes their own connections and connections shared with their workspaces, and the web search, AI model, social-media and Agent Email connections every account includes; does not connect an app or return credentials. Returns each connection's name, status, errors and sharing; scriptit integrations list without --brief adds its scopes and API docs. Returns one page of integrations with their total and a next_cursor; pass it as cursor for the next page. For general work or additional options, use the documented commands through run_command. Requires an existing Script.it session_id. Runs as a tracked command in that session. Read output and exit_code; a running command is not a completed operation. Continue pending or paginated output with read_output using the returned session_id, command_id and next_offset; do not repeat the operation to poll it.

list_triggers

ChatGPT
Command shortcut: scriptit trigger list --brief. Read 'scriptit skills show triggers --file MANAGE.md' through run_command for the full workflow, examples and command options. Inspect the user's configured schedules and event triggers, including their settings, target scripts and enabled or paused state. Check for an existing trigger before creating another, and verify the state after enabling or disabling one. Returns each trigger's identity, state, settings and target script; scriptit trigger get <id> shows one in full. Returns one page of triggers with their total and a next_cursor; pass it as cursor for the next page. For general work or additional options, use the documented commands through run_command. Requires an existing Script.it session_id. Runs as a tracked command in that session. Read output and exit_code; a running command is not a completed operation. Continue pending or paginated output with read_output using the returned session_id, command_id and next_offset; do not repeat the operation to poll it.

read_output

ChatGPT
Wait for a command run_command started and read its output from offset bytes on. While the command runs, a call waits about 40 seconds for it to end. next_offset is null once the output has been read to its end.

run_command

ChatGPT
Write files into a Script.it session that start_session opened, then run a bash command there; either may be left out. Files are written exactly as given, with no shell in between, replacing any file already at the path, at paths relative to the session's directory: workspaces/<workspace-id>/... for the user's scripts, data_files/... for the session's own files. The command runs in the session's directory, where scriptit is the Script.it runtime: running scripts, listing and showing skills, and the other verbs the session's context describes. A call returns within about 40 seconds; a command still running by then comes back with status running, and read_output follows it. Long output comes in pages that read_output continues from next_offset.

run_script

ChatGPT
Command shortcut: scriptit start <path> --session new [--<input-name> <value> ...]. Read 'scriptit skills show script-authoring' through run_command for the full workflow, examples and command options. Run a saved workspace script with typed inputs, for a one-time task or a test before setting up recurring or event-triggered work. Starts the workflow in a child session and returns its run_id and session_id without waiting for the workflow to finish. Use that child session_id with get_run_status, and with scriptit logs through run_command. The script must have a stable ID. Its steps can write files and act in connected apps or on external websites; run it only for the user's requested task. For general work or additional options, use the documented commands through run_command. Requires an existing Script.it session_id. Runs as a tracked command in that session. Read output and exit_code; a running command is not a completed operation. Continue pending or paginated output with read_output using the returned session_id, command_id and next_offset; do not repeat the operation to poll it.

start_session

ChatGPT
Open a session in the user's Script.it sandbox, the cloud environment holding their skills, scripts and connected integrations. Returns the session's session_id and its whole context: how Script.it works, which integrations are connected and which skills exist. Work in the session relies on all of the context, and goes through run_command.

validate_script

ChatGPT
Command shortcut: scriptit validate <path> --json-output --agent-diagnostics [--fix]. Read 'scriptit skills show script-authoring' through run_command for the full workflow, examples and command options. Check an authored automation's schema, source metadata and diagnostics before running it. Does not execute its workflow steps. With fix_metadata, also assign a missing stable script ID and repair generated cover metadata and trigger declarations. For general work or additional options, use the documented commands through run_command. Requires an existing Script.it session_id. Runs as a tracked command in that session. Read output and exit_code; a running command is not a completed operation. Continue pending or paginated output with read_output using the returned session_id, command_id and next_offset; do not repeat the operation to poll it.

whoami

ChatGPT
Show which Script.it account this connection acts for. Work in Script.it itself happens in a session start_session opens.

Capabilities

Writes

Example Prompts

Click any prompt to copy it.

Similar Apps

Listed on MCP App Store

Listed on MCP App Store

Add this badge to your site or README. It links to this page.

HTML

<a href="https://mcp-app-store.keenethics-labs.com/app/script-it"><img src="https://mcp-app-store.keenethics-labs.com/badges/listed-on-mcp-app-store.svg" alt="Listed on MCP App Store" width="220" height="44"></a>

Markdown

[![Listed on MCP App Store](https://mcp-app-store.keenethics-labs.com/badges/listed-on-mcp-app-store.svg)](https://mcp-app-store.keenethics-labs.com/app/script-it)