Skip to content
OpenAgentSkill
Browsing is open · Catalog captured 2026-11-21 · Automatic updates and account services are paused.
Development

powershell-write-tests

Write or extend Pester 6 tests for existing PowerShell code, following this repository's testing rules. Use when asked to add tests, raise coverage, or test a PowerShell function, script, or module.

fadwen · Static checks on file

Installation

Read the source and dependencies before running a command. Copying does not install or execute anything.

npx skills add fadwen/ai-powershell-standards --skill powershell-write-tests

About this skill


name: powershell-write-tests description: Write or extend Pester 6 tests for existing PowerShell code, following this repository's testing rules. Use when asked to add tests, raise coverage, or test a PowerShell function, script, or module.

Write Pester tests for PowerShell code

Read .github/instructions/powershell-pester.instructions.md first, unless it is already loaded. It holds the rules, a complete test file, and a table of assertion behaviour.

Steps

  1. Read the code under test completely: its parameters and parameter sets, validation attributes, output type, every catch, and every call that leaves the process (network, files, other modules, the clock).
  2. Look at an existing test file in the repository. Keep its assertion style and layout. In a new file use the Pester 6 Should-* assertions.
  3. List the behaviours to test before writing any: each parameter set, the documented output, each validation failure, each error path, pipeline input, and for a Test- function both outcomes.
  4. Write the file so that it runs on its own: a versioned #Requires for Pester, the code loaded in a top-level BeforeAll, and test data built in BeforeDiscovery.
  5. Mock only what crosses a boundary. For code inside a module, pass -ModuleName to Mock and Should-Invoke, and reach private functions with InModuleScope.
  6. Make every test able to fail: assert specific values, assert the count before reading by index, and assert expected errors with Should-Throw or by counting error records.
  7. Run the tests. A test that passes on its first run has not been shown to work: break the behaviour it covers, or the expected value, once to see it fail, then restore it.

Report

Say which behaviours are covered and which are not, the pass and fail counts, and the coverage figure if you measured it. If a test exposes a defect in the code, report the defect; do not weaken the test to make it pass, and do not change the code under test unless the user asked for that.