Armature Launches Agent.reviews, A Review Site Where AI Agents Rate The Tools They Use
Agent.reviews turns the reviews AI agents write about tools into a selection mechanism aimed at other agents, which puts an evaluator role in the same software category it is rating.
Reporting from 1 source: GIGAZINE.
Armature, an AI-related company, developed agent.reviews, a review site where AI agents post evaluations of software and development tools they have used. Reviews cover three categories, usefulness, ease of use, and reliability, each scored on a 5-point scale, and record the work performed and how the tool was connected. The OpenAI API page showed 1,749 reviews and an overall rating of 4.2 at the time of writing. Reviews can be filtered by the agent that posted them, including Claude Code, Codex, and Cursor.
Reviews are scored in three categories: usefulness, ease of use, and reliability, each on a 5-point scale. A review also records the work the agent actually performed, how the tool was connected, and where problems occurred, so the score is tied to a specific task rather than a general impression. The site gives the OpenAI API as an example: 1,749 reviews and an overall rating of 4.2 at the time of writing.
Agents post through an "agent-review" skill that can be installed by running a prompt, and initial settings have the agent review the tools it used after each completed task. Rankings favor tools that have collected 5 or more reviews.
Synthesized by Yomimono from the 1 cited source below, including Japanese-language reporting where cited, then editorially reviewed before publishing.