Testing & QA · By dxos

Record a demo of the running app that the agent drives itself — a .mdl QA test or an ad-hoc walkthrough — as a captioned .webm (or a screenshot), trimmed of dead air and ready to attach.

dxos/dxos

Authoring browser end-to-end tests for the Composer app with Playwright.

dxos/dxos

Working inside the Claude Code cloud sandbox (Claude Code on the web, remote sessions, scheduled runs).

dxos/dxos

Land an existing PR — finds it, fixes CI failures iteratively, keeps the branch up to date with main, subscribes to PR events for continuous autofixing, and adds to merge queue.

dxos/dxos

Regenerate cached LLM responses for tests that replay model fixtures.

dxos/dxos

CPU-profile or leak-detect an existing vitest suite via env-gated instrumentation — DXPROFILETESTS emits a V8 .cpuprofile, DXDEBUGLEAKS writes before/after heap snapshots plus a per-test heapUsed…

dxos/dxos

Test assistant conversations, agents, and skills using AssistantTestLayer, Effect/vitest, ECHO types, and memoized LLM fixtures.

dxos/dxos

Use Trunk's CI Autopilot MCP server to investigate a failing CI run, find/diagnose a flaky test, and quarantine it.

dxos/dxos