RequestTailor Preview (22 November 2026): requests as files in Git, no account, offline – on the Mac, in the browser and in the pipeline. What’s inside → +43 664 1964199Contact →

AI agents and MCP server

rt mcp offers a collection to coding agents as tools over the Model Context Protocol. The agent sends requests exactly as you would – from the same files, with the same core. You decide per environment what it may do. The default is read-only; a writing request against production waits until you approve it. The agent never sees secret values, and every call goes into a local audit log.

Edition
Free
Start
rt mcp <collection> (stdio)
Default
read-only
Approval
banner in the app or rt approve/rt deny
Audit log
agents/audit.jsonl in the data folder, local only

Connect an agent

rt mcp speaks MCP over stdin and stdout (JSON-RPC). The MCP client starts the process itself; you only give it the command. In Claude Code:

claude mcp add requesttailor -- rt mcp /path/to/collection

Other clients such as Cursor, VS Code or Claude Desktop take the same command in their JSON configuration:

{
  "mcpServers": {
    "requesttailor": {
      "command": "rt",
      "args": ["mcp", "/path/to/collection", "--env-default", "DEV"]
    }
  }
}
OptionMeaning
--var NAME=VALUEVariable for every send, such as a token; treated as a secret – the agent only sees *** and may not override it
--env-default NAMEEnvironment when the agent names none
--approval-timeout SECONDSHow long a request waits for approval (default 300); then it fails and nothing is sent
--no-scriptsDo not run scripts
--agent-name NAMEName for approvals and the audit log when the client sends none
--body-limit BYTESHow much of the response body the agent gets per request (default 64 KiB)

The tools

ToolKindWhat it does
list_requestsread-onlyPath, name, method and effect (read or write) of every request
get_requestread-onlyOne request as the model sees it; credentials masked
list_environmentsread-onlyThe environments the agent may use, with access, production flag and non-secret values
send_requestdestructiveSends one request and returns the masked report
run_folderdestructiveRuns a folder or the collection like rt run; every request is checked on its own

Every call reads the files anew – the agent always works with the current state of the repository.

Access per environment

What agents may do belongs to the collection and lives in opencollection.yml. In the app you set it in the Agent access dialog (command palette: Agent access…).

Agent access dialog of the collection Demo API: Default for this collection Read and write, environment Local with Default, environment PRD with Read and write and Production ticked, column Agents may with Writing requests wait for approval; an approval banner above
Local may read and write, PRD is marked as production: writing requests wait for approval there.
  1. Default for this collection applies to every environment without a rule of its own.
  2. Per environment, Access: no access, read only, or read and write.
  3. Production marks a production environment. If an environment is named like production but not marked, the dialog points it out.
  4. Agents may sums up what follows. It is saved with the collection (Save).
extensions:
  requesttailor:
    agents:
      access: read                 # default: none, read or write
      environments:
        DEV: { access: write }
        PRD: { access: write, production: true }
        HR: none                   # hidden from agents
  • none: agents do not see the environment and cannot send in it.
  • read: only reading requests go out, writing ones are refused. This is the default.
  • write: writing requests go out – except in an environment with production: true. There every writing request waits for your approval.

Reading or writing?

GET, HEAD, OPTIONS and TRACE read, everything else writes. GraphQL reads as long as the document contains no mutation; gRPC writes; a WebSocket writes as soon as it sends messages. list_requests shows the agent this classification up front.

Approvals

A writing request against a production environment writes an approval request to the data folder and waits. You answer in one of three places; the first answer wins:

Red banner above the request editor: claude-code wants to send POST https://api.example.com/users?api_key=**** in PRD, below it collection, file users/create-user.yml and writing request against the production environment PRD, with Approve and Deny on the right
The approval banner appears on every screen of the app; the API key in the URL is masked.
  • in the Mac app: a banner on every screen with method, masked URL, environment and agent, with Approve and Deny;
  • on the command line: rt approvals lists what is waiting (--json for scripts), rt approve <id> and rt deny <id> answer;
  • in the MCP client itself, if it supports MCP elicitation (a yes/no question). The client’s answer counts as yours – only use this with clients that ask a person, not the model.

Denied, timed out or cancelled: the request is not sent, the agent gets an error that says so, and the approval request is removed.

The audit log

Every tool call and the start of every session append one JSON line to agents/audit.jsonl in the data folder: time, agent, tool, collection, request, environment, method, masked URL, status, outcome (read, sent, refused, denied, timed out, failed), who approved (app, command line, elicitation) and a detail. In the app, Agent audit log… in the command palette shows it. The log stays on this computer; nothing is sent anywhere.

What limits the agent

  • No secret values. The core masks reports already. On top of that, rt mcp replaces every secret value the process knows with *** in every message to the agent and every audit line – also when a server echoes a token back in its body. get_request and list_environments leave secret values out.
  • Rules pinned at start. An agent that can edit files could otherwise give itself write access. Later changes apply only where they are stricter; loosening needs a restart.
  • Agent variables redirect nothing. They may override neither variables of the collection, the environment or --var (base URLs and credentials live there) nor any secret name.
  • OAuth sign-ins that need a browser or a code are not started by rt mcp; it uses a token the app has already fetched.
  • Scripts run only in collections you trusted in the Mac app, and not with --no-scripts. Requests a script sends itself are not checked against the rules – do not trust collections whose scripts you do not know.

More on the security model on the Security page.

See also

Sources on this page

The superscript numbers in the text point here. Tap one to see the source right where it is used; all sources of this website are listed on one page.

  1. 1

    MCP server rt mcp with rules per environment ↑Product · RequestTailor core repository, crates/rt-cli/src/mcp, crates/rt-cli/README.md (MCP server), crates/rt-cli/tests/gate_w8.rs · checked on 2026-11-22

  2. 2

    No account, offline, no connection to CloudDNA ↑Product · RequestTailor core repository, CLAUDE.md (Produktprinzipien 2 und 7), website/docs/privacy · checked on 2026-11-22

Frequently asked

Short answers.

Which AI agents work with RequestTailor?
Any MCP client that starts a server over stdio, such as Claude Code, Cursor, VS Code or Claude Desktop. rt mcp supports the MCP protocol versions 2025-06-18, 2025-03-26 and 2024-11-05.
May an agent send writing requests without any setup?
No. Without a rule, every environment is read-only; writing requests are refused. Only with write may it write, and in an environment marked as production every writing request waits for your approval.
What happens if I do not answer an approval?
After the timeout (default 300 seconds, set with --approval-timeout) the request fails and nothing is sent. The agent gets an error message that says so.
Is the audit log sent to CloudDNA?
No. It is a file, agents/audit.jsonl, in the data folder on your computer. RequestTailor opens no connection to CloudDNA servers.

Your question isn’t here? Talk to me: +43 664 1964199 office@clouddna.at Get in touch

Contact

Requests in the repo instead of an account? Talk to us.

Free is free and stays free. For an early copy, for Enterprise, for the OData plugin on your SAP system or for a wish for Pro and Team: write to me. I reply within one working day, personally.

Martin KochManaging Director, CloudDNAI read every request myself and reply personally.