Paired-Twin Change Trials

com.textforagentsPublished by textforagents.comOfficial publisherv1.0.0Updated Oct 26, 2026

Checkpoint-paired twin trials for agent fleet changes with blinded verdicts.

The publisher's own server: Agent NOTAMs: hazard bulletins for agents

VerifiedStreamable HTTPWeb executableData & Analytics AI & ML

Server details

Identifier: com.textforagents/paired-twin-change-trials-for-agents
Transport: Streamable HTTP
Runtime: Web executable
Trust: Verified
Published: Oct 22, 2026

Security note

MCP servers receive tool arguments from a conversation and may perform external actions. Review what a server can do before you enable it, and give it only the access it needs.

The registry verified that this server's namespace belongs to textforagents.com, which also runs its endpoint. That identifies the publisher; it does not make the server's tools safe.

SourceWeft's checks cover whether the listing is complete and the endpoint answered; they do not review the server's code or what its tools do.

Add a badge to your README

Show readers this MCP server is listed on SourceWeft. The badge links to this page.

[![Listed on SourceWeft](https://sourceweft.com/badge/mcp/com.textforagents%2Fpaired-twin-change-trials-for-agents.svg)](https://sourceweft.com/mcp/com.textforagents%2Fpaired-twin-change-trials-for-agents?utm_source=readme&utm_medium=badge)

Overview

AI-generated overview

Remote MCP server for running blinded, checkpoint-paired A/B trials of agent fleet changes, with a stopping-rule verdict and signed receipt.

What you can do: Register a hash-bound trial protocol where twins fork from an identical checkpoint and only one registered variable differs
Fork checkpoint-paired, blinded A/B arms for a task
Run a pair with a deterministic scripted agent and score it blind, or submit your own twins' outputs for blinded scoring
Get the trial verdict under the enforced stopping rule
Unblind after the verdict and receive a hashed signed receipt covering every pair
When to use it: Useful when you want to test whether a single change to an agent fleet actually helps, using paired twins from the same checkpoint and blinded scoring rather than an informal comparison. Suited to teams that need a documented verdict and a signed receipt for the result.
Requirements: A remote streamable HTTP endpoint at textforagents.com; no packages, environment variables or headers are declared. Paid per pair beyond the free allowance, so an operator account and payment arrangement are implied.

Before you install

Running or submitting pairs costs $0.25 per pair after the first 10 free pairs per operator, so usage spends money. Trials write registered protocols and submitted outputs to the remote service, and results stay blinded until the verdict is issued.

Installation

In SourceWeft

Open Paired-Twin Change Trials in the dashboard and add it to a workspace.
Enable the server for the chats that should use its tools.

Web executable via Streamable HTTP. Remote servers run from the web runtime once configured in a workspace.

Use with an agent

Paste this prompt into Claude Code, Codex, Cursor or another agent — or into SourceWeft chat. The agent reads this server's install guide, shows you its publisher, endpoint and sign-in, and adds it to your client once you agree. Keys never go into a command.

Read https://sourceweft.com/mcp/com.textforagents%2Fpaired-twin-change-trials-for-agents/install.md and connect the MCP server it describes to my client. Before changing any config, show me its publisher, endpoint and sign-in method and wait for my OK. Never put an API key in a command or in a file I would commit; ask me where it should come from.

Read the install guide the agent follows

Other MCP clients

Add this to your client's mcpServers config.

{
  "mcpServers": {
    "paired-twin-change-trials": {
      "type": "http",
      "url": "https://textforagents.com/products/paired-twin-change-trials-for-agents/mcp"
    }
  }
}

Tools

trial_register_protocol

Register a hash-bound trial protocol; twins fork from the identical checkpoint, exactly one registered variable may differ.

Parametersprotocolrequiredcheckpointrequiredtreated_configrequired

trial_fork_pair

Fork a checkpoint-paired assignment for one task (blinded arms A/B).

Parameterstaskrequiredtrial_idrequired

trial_run_pair

Run one pair with the deterministic scripted agent and score blind. $0.25/pair, first 10 free once per operator.

Parameterstaskrequiredtrial_idrequired

trial_submit_outputs

Submit the operator twins' outputs for one pair; blinded deterministic scoring. $0.25/pair, first 10 free once per operator.

Parameterstaskrequiredtrial_idrequiredcontrol_outputrequiredtreated_outputrequired

trial_score

Alias of trial_submit_outputs (blinded scoring).

Parameterstaskrequiredtrial_idrequiredcontrol_outputrequiredtreated_outputrequired

trial_verdict

Trial verdict under the enforced stopping rule.

Parameterstrial_idrequired

trial_unblind

Unblind after the verdict; hashed signed receipt covering every pair.

Parameterstrial_idrequired

Version history

v1.0.0LatestOct 22, 2026

Score

83% of the applicable checks met

How complete this listing is, computed from the registry entry, the endpoint and public sources. Not a rating of the server and not a statement that it is safe or working today: reachability is a dated check, and no installs or reviews go into the score.

Reached in the last 30 daysRequired25 pts
Auth declaredRequired10 pts
Tools listedRequired15 pts
DescriptionRequired15 pts
Logo10 pts
Documentation15 pts
Declare a GitHub repository with a README, or serve resource_documentation in the OAuth protected-resource metadata.
LicenseNot applicable
A hosted service is used under its provider's terms, so no license is expected here. This check is left out of the percentage.