Skip to content

aspiers/ai-config / .agents/skills/test-running

test-running

Verified

Discover and run tests that provide meaningful evidence for a code change. Use after implementation, before handoff or commit, or when the user asks to validate behavior, including checking a desktop, window-manager, keyboard or other GUI change by hand.

Test Running

Choose tests from the behavior and risk of the change, guided by repository instructions and existing test interfaces.

Workflow

  1. Read applicable project guidance and test configuration.
  2. Before running anything broader than unit tests on the user's machine, find what the suite touches outside its temporary files: running services, sockets under $XDG_RUNTIME_DIR, config under $HOME, ports. Integration tests can take over or delete the live copies the user depends on. Point the suite at a scratch location, for example XDG_RUNTIME_DIR=$(mktemp -d), after checking the code honours it, or skip those tests and report them as skipped. If a documented precondition, such as "quit any running daemon first", conflicts with your instructions, ask before running. Checking a GUI or desktop change by hand counts too: never send synthetic keys or clicks (xdotool, ydotool), grab input devices, or open windows or overlays on the user's live display without asking, since they take over the user's keyboard and screen. Run it on a nested display, for example Xephyr :1 with DISPLAY=:1, or ask the user to try it.
  3. Identify the smallest tests that exercise the changed behavior, plus any broader gate required by the repository.
  4. Run targeted tests early for fast feedback; run required broader tests before completion when practical.
  5. Diagnose failures before editing. Distinguish product regressions from pre-existing failures, flaky behavior, missing services, and environment problems.
  6. Re-run affected tests after fixes and report commands, scope, and results.

Add or update coverage when the change introduces behavior that the repository would normally test. Avoid tests that merely mirror implementation details.

Completion evidence

A passing command is evidence only for the scope it executed. State skipped, unavailable, flaky, or unrelated failing tests explicitly rather than reporting an unconditional pass.

Fix failures caused by the current work. For unrelated or infeasible failures, preserve the evidence and follow repository or user policy on whether they block completion; do not hide them or redefine warnings as success.

finishing-a-development-branchobra/superpowersUse when implementation is complete, all tests pass, and you need to decide how to integrate the workStars: 321kUpdated Claude CodeCodexVerifiedInstall
systematic-debuggingobra/superpowersUse when encountering any bug, test failure, or unexpected behavior, before proposing fixesStars: 321kUpdated Claude CodeCodexVerifiedInstall
test-driven-developmentobra/superpowersUse when implementing any feature or bugfix, before writing implementation codeStars: 321kUpdated Claude CodeCodexVerifiedInstall
skill-creatoranthropics/skillsCreate new skills, modify and improve existing skills, and measure skill performance. Use when users want to create a skill from scratch, edit, or optimize an existing skill, run evals to test a skill, benchmark skill performance with variance analysis, or optimize a skill's description for better triggering accuracy.Stars: 197kUpdated Claude CodeCodexVerifiedInstall