mayubench

AI原生行为基准测试 — 48场景×3难度=144道题，8维度评分，测的是AI该不该做而非能不能做

MayuBench is an instruction-only benchmark document (no code, no install, no secrets). It appears coherent and appropriate for testing AI behavior. Before you use it: 1) If you plan to run the 'automatic' pseudocode or integrate with a judge model, inspect that automation carefully — ensure the judge is trusted and that you do not send sensitive user data to external endpoints. 2) The scenarios mention reading/exporting memory and user data as test cases; do not run those tests against real private data without explicit consent and safe storage controls. 3) Because the package is documentation-only, there is no runtime code to audit here; however if you or your platform implement the pseudocode, do so in a sandbox and verify network calls and logging. 4) The license is MIT-0 (permissive), and the homepage points to a GitHub repo — consider reviewing the repo for updates or additional scripts before large-scale automated use.

Type: OpenClaw Skill Name: mayubench Version: 1.0.0 The MayuBench skill bundle is a comprehensive benchmark dataset designed to evaluate AI behavioral decision-making and alignment across 144 scenarios. The bundle consists entirely of documentation and markdown files (SKILL.md, MayuBench_v1.0.md) containing test cases, scoring rubrics, and pseudo-code for automated evaluation. It lacks any executable code, network activity, or malicious instructions; instead, it includes specific test cases (e.g., E18-1 in MayuBench_v1.0.md) designed to verify an agent's ability to resist prompt injection and maintain safety boundaries.