documents for agents
This repository also configures its own agents. See what pinqdoq tells them →
Agent Claude Code
Runs Maestro flows on the Android emulator / iOS simulator for a pinqloq task runner task, inspects the screenshots, optionally runs the pinqloq log queries for the run window, and returns a short verdict with evidence paths. Used so screenshots, hierarchy dumps and device logs never enter the runner's own context.
Agent Claude Code
Runs the pinqcode-review skill on a pinqloq task runner branch (origin/ ...HEAD) and returns only the findings list. Used so the full review transcript never enters the runner's own context.
Agent Claude Code
Runs builds, unit/UI/screenshot tests, PIT or Stryker.NET for a pinqloq task runner worktree and returns only a short verdict. Used by the pinqloq-task-runner so long Gradle/dotnet/mutation output never enters the runner's own context.