[rohanb123]

Description

#5310 made == and != coerce numeric strings the same way >, <, >= and <= already did, so captured stdout such as "10" is now == 10. Its regression test used a hand-built StepContext with stdout: "10". Real shell output keeps its trailing newline (echo 10 captures "10\n"), and no test covered that path end to end.

This adds one parametrized test to tests/workflows/test_shell_step.py, written like the existing engine-level tests in that file. A shell step prints a value and an if step branches on steps.count.output.stdout. The test checks:

• == 10 / != 11 take then, and != 10 / == 11 take else, for stdout "10\n"
• >= 10 and <= 10 agree with == on the same value
• a decimal value ("2.5\n" == 2.5) compares as a number
• non-numeric output ("10x\n") is still compared as a string, so == 10 is false and != 10 is true
• the raw stdout field still keeps its trailing newline

This only adds tests. Runtime behavior is unchanged.

Testing

Set up this checkout's .venv and editable install using the local development guide. Run commands from the checkout root.

Ran in a clean python:3.14 container:

pip install --no-cache-dir uv && uv sync --extra test
export LC_ALL=C.UTF-8 && .venv/bin/python -m pytest tests/workflows/test_shell_step.py -q && uvx ruff@0.15.0 check tests/workflows/test_shell_step.py

Result: pytest collected 160 items in tests/workflows/test_shell_step.py and reported 160 passed in 1.12s. ruff check on the same file reported All checks passed!. The command exited with status 0.

• Tested locally with .venv/bin/specify --help (Windows PowerShell: .\.venv\Scripts\specify.exe --help)
• Ran relevant existing tests with .venv/bin/python -m pytest tests -q (Windows PowerShell: .\.venv\Scripts\python.exe -m pytest tests -q); use affected test paths for focused checks and LC_ALL=en_US.UTF-8 on macOS
• Tested with a sample project (if applicable)

AI Disclosure

• I did not use AI assistance for this contribution
• I did use AI assistance (fill in the disclosure below)

AI disclosure: Written by Claude Code (model: Claude Opus 5.5, low reasoning effort, autonomous) on behalf of Rohan Bharti. The agent picked the gap, wrote the test and wrote this description. Nobody reviewed the change line by line before it was opened.

---

Prepared with AI assistance (Claude Code). The tests described above were run before opening this PR. The submitter reviewed the diff before it was opened.

[Copilot]

🟢 Approval recommended

The review records no active issues, and the change is limited to focused regression coverage for shell-output comparisons.

No open issues 🎉

What changed in this PR

Adds end-to-end regression coverage for numeric comparisons against shell-step stdout with a trailing newline. Runtime behavior is unchanged.

Changes:

• Tests equality, inequality, and ordering comparisons against captured stdout.
• Covers decimal values and non-numeric string fallback behavior.
• Verifies raw stdout retains its trailing newline and conditional branches behave correctly.

File	Description
tests/​workflows/​test_shell_step.py	Adds parametrized engine-level coverage for numeric and string comparisons on captured shell output.

_{Review effort: Lite (auto)}

[Copilot]

🟢 Approval recommended

No active review issues remain, and the changes are limited to focused end-to-end regression coverage for numeric comparisons involving shell-step stdout.

No open issues 🎉

_{Review effort: Lite (auto)}

[mnriem]

Thank you!