fix: ensure bash scripts are executable out of the box (#11)

[raza-khan0108]

Fix: Ensure bash scripts in templates are executable out-of-the-box (#11)

Summary

This PR resolves Issue #11 ([Bug]: Bash script not executable ootb), where AI agents and developers encountered permission errors such as:

(eval):1: permission denied: .specify/scripts/bash/setup-plan.sh

When project templates were initialized or extracted, script files often lacked POSIX execute permissions (+x) out-of-the-box (OOTB). This PR fixes the issue across three distinct layers: git tracking index permissions, GitHub Actions template release packaging, and the Specify CLI runtime post-processing.

---

Key Changes

1. Git Repository Tracking Permissions

Normalized git index file modes from 100644 (read/write non-executable) to 100755 (executable) for all bash scripts in scripts/bash/:

• scripts/bash/check-prerequisites.sh
• scripts/bash/common.sh
• scripts/bash/create-new-feature.sh
• scripts/bash/setup-plan.sh
• scripts/bash/update-agent-context.sh

2. Release Archive Packaging Workflow

Updated .github/workflows/scripts/create-release-packages.sh to guarantee that any downloaded .zip template archive from GitHub Releases has executable permissions pre-configured:

• Added an explicit normalization command right before creating each zip package in build_variant():

  find "$base_dir" -type f -name "*.sh" -exec chmod +x {} + 2>/dev/null || true

3. Specify CLI Core (__init__.py)

Enhanced ensure_executable_scripts() in src/iac_specify_cli/__init__.py:

• Expanded Scope: Recursively searches and sets POSIX execute bits on all .sh scripts under both .specify and .speckit project directories (rather than strictly .specify/scripts).
• Windows Tracker Fix: On Windows (os.name == "nt"), ensure_executable_scripts() now properly registers the chmod step as skipped in the Rich StepTracker. This prevents the progress spinner from hanging or remaining in an incomplete state during iac-specify init on Windows machines.

4. Test Suite Compatibility & Unit Tests

Updated tests/test_ensure_functions.py:

• Added @pytest.mark.skipif(os.name == "nt", ...) decorators to POSIX permission assertion tests that cannot execute natively on Windows NTFS.
• Added a new unit test (test_makes_sh_scripts_in_speckit_executable) to verify that scripts inside .speckit/ are made executable.
• Updated test_tracker_reports_updated_count to accept "skipped" status on Windows and "done" on POSIX systems.

5. Versioning & Documentation

In accordance with AGENTS.md guidelines for changes touching __init__.py:

• Bumped CLI version from 0.0.9 to 0.0.10 in pyproject.toml.
• Added complete release notes under ## [0.0.10] - 2026-07-07 in CHANGELOG.md.

---

Related Issue

Closes #11

---

Verification & Testing

Git Mode Verification

Verified git index mode updates via git diff --cached --summary:

 mode change 100644 => 100755 scripts/bash/check-prerequisites.sh
 mode change 100644 => 100755 scripts/bash/common.sh
 mode change 100644 => 100755 scripts/bash/create-new-feature.sh
 mode change 100644 => 100755 scripts/bash/setup-plan.sh
 mode change 100644 => 100755 scripts/bash/update-agent-context.sh

Automated Unit Tests

Ran the full test suite across Windows and POSIX environments:

$ uv run pytest
======================= 330 passed, 3 skipped in 1.73s ========================

• 0 failures, 330 passed, 3 skipped (POSIX-specific tests cleanly skipped on Windows).

---

Checklist

• I have searched existing issues for duplicates.
• My code follows the style guidelines and general practices in AGENTS.md.
• I have updated the package version in pyproject.toml and documented changes in CHANGELOG.md.
• I have added/updated unit tests verifying the fix across both POSIX and Windows platforms.
• All automated unit tests pass locally.

[Copilot]

Copilot was unable to review this pull request because the user who requested the review has reached their quota limit.