agent-qa

Vostride's source-available agentic QA harness for web, Android, and iOS: teams write reviewable natural-language tests, run them through a dashboard, CLI, or MCP server, and retain artifacts plus file-backed memory between runs. It ships three Agent Skills for authoring, failure triage, and evidence-led debug/fix work. The honest catches: model-backed runs still cost inference and can be nondeterministic; browser/mobile setup is substantial; and FSL-1.1-ALv2 restricts competing commercial use until each release converts to Apache-2.0 two years later.