Sign inGet started
–

Audits and prunes redundant, vacuous, or implementation-coupled tests while preserving meaningful regression coverage to reduce maintenance and CI costs.

Lift · GLM-5.3-flash+25ppdirectional · all models →
88
Trustsecurity · quality · license · upkeep
Scan Saferisk 0/100
LicenseMITserved via MCP
Not yet benchmarked

New and updated skills run on the daily bench queue: same container, same tools, with and without the skill — first on the reference model, then on Claude Haiku 5.5 and GPT-6 Luna when there’s a signal. How benchmarking works →