fix(workflows): reject a non-string 'command' in command-step

[Noor-ul-ain001]

Summary

CommandStep.validate only checked that a command field is present, never its type. But the workflow engine does not auto-validate a step before calling execute, so on an unvalidated run a non-string command — null, a list, an int — was passed straight through _try_dispatch to the integration's build_command_invocation, which does:

stem = command_name
if stem.startswith("speckit."):   # AttributeError if command_name isn't a str

Once a resolvable integration with an installed CLI is found, this crashes the entire workflow run with a raw AttributeError instead of failing just the step.

Confirmed directly:

['a', 'b'] -> AttributeError 'list' object has no attribute 'startswith'
None       -> AttributeError 'NoneType' object has no attribute 'startswith'
42         -> AttributeError 'int' object has no attribute 'startswith'

This is the same class of gap that #3582 closed for the prompt-step prompt field — validate() checked presence only while execute() fed the value to the CLI layer.

Fix

Guard both paths, matching the sibling steps already in this file:

• validate() rejects a non-string command, mirroring prompt-step prompt (fix(workflows): reject a non-string prompt in prompt-step validate() #3582) and shell-step run.
• execute() fails the step cleanly with the same contract error before dispatch, mirroring the existing input/options guards in CommandStep.execute — so an unvalidated run FAILs the step (honouring continue_on_error) instead of crashing the run.

An expression like {{ inputs.cmd }} is still a str, so it stays valid.

Tests

Added to TestCommandStep:

• test_validate_rejects_non_string_command (null / list / int / dict, plus accepts a plain string and an expression)
• test_execute_non_string_command_fails_cleanly (patches shutil.which so dispatch would actually be attempted absent the guard — proving the guard prevents the crash)

Verified test-the-test: with the source fix stashed, validate() misses the bad value and execute() crashes at build_command_invocation with the raw AttributeError; with the fix, all 13 TestCommandStep tests pass. The 20 unrelated test_workflows.py failures are the pre-existing Windows symlink-guard tests that require elevation.

🤖 Generated with Claude Code

[Copilot]

Pull request overview

Adds robust type validation for workflow command steps, preventing non-string commands from crashing dispatch.

Changes:

• Rejects non-string command values during validation and execution.
• Adds coverage for validation and clean step failure behavior.

Show a summary per file

File	Description
src/specify_cli/workflows/steps/command/__init__.py	Adds command type guards.
tests/test_workflows.py	Tests invalid command handling.

Review details

Tip

Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

• Files reviewed: 2/2 changed files
• Comments generated: 0
• Review effort level: Medium

[mnriem]

Thank you!