Agents

Executes the mat-agent-bench benchmark against a local server, fetching questions, executing tasks, and submitting results for grading.

Install this skill

99/100

Security score

The Agents skill was audited on May 23, 2026 and we found 1 security issue across 1 threat category. Review the findings below before installing.

low line 83

External URL reference

SourceSKILL.md

83	- The server URL is provided when this skill is invoked (e.g., `http://127.0.0.1:8765`)

Scanned on May 23, 2026

GitHub Stars 1

Rate this skill

Categoryproductivity

UpdatedMay 31, 2026