fix: simplify Claude workflows and TestPyPI publishing

[spazyCZ]

Summary

• simplify the Claude review workflows to use comment-only prompts instead of the old plugin-based flow
• fix the TestPyPI publish workflow to use unique dev versions on the test branch and Trusted Publishing via the testpypi environment
• update .gitignore to a standard Python baseline including coverage artifacts

Included changes

• replace the old Claude Code Review plugin/slash-command configuration with a plain review prompt
• keep Claude Quality Gate on the comment-only review path
• add automatic TestPyPI version stamping before build using a unique .dev version per run
• set environment: testpypi and id-token: write for Trusted Publishing
• expand .gitignore for common Python, packaging, test, and coverage artifacts

Verification

• workflow YAML files validated locally with yaml.safe_load
• branch pushed successfully

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ All tests successful. No failed tests found.

📢 Thoughts on this report? Let us know!

[Copilot]

Pull request overview

This PR contains three main categories of changes: simplification of Claude AI-assisted CI workflows (removing plugin-based configuration in favor of plain prompts), fixing the TestPyPI publishing workflow to use Trusted Publishing via environment-based OIDC tokens (replacing API key auth), and a significant refactoring of ptop3/monitor.py including removal of the non-functional net sort key, encapsulation of sampling state in a ProcessSampler dataclass, addition of scroll support to the TUI, and an extensive new test suite.

Changes:

• Claude workflows (claude-code-review.yml, claude-quality-gate.yml) simplified to use plain review prompts and restricted to comment-only mode (no auto-commits).
• TestPyPI workflow upgraded to Trusted Publishing (id-token: write, environment: testpypi) with automated unique dev-version stamping per run.
• ptop3/monitor.py refactored: net sort key and net I/O fields removed; process sampling encapsulated into ProcessSampler dataclass; scroll support added to TUI via visible_window_start/_sync_scroll; helper functions extracted (sort_groups, sort_processes, _selected_group, _selected_proc, _status_line); large test suite added covering all new/changed functions.

Reviewed changes

Copilot reviewed 11 out of 12 changed files in this pull request and generated 1 comment.

Show a summary per file

File	Description
.github/workflows/publish-testpypi.yml	Switched to Trusted Publishing; added unique dev-version stamping step
.github/workflows/claude-code-review.yml	Replaced plugin/slash-command with plain review prompt; widened permissions to allow PR comments
.github/workflows/claude-quality-gate.yml	Agents restricted to comment-only (no auto-commits); added test branch trigger
ptop3/monitor.py	Major refactor: ProcessSampler dataclass, net sort removal, scroll support, helper extraction
tests/test_monitor.py	~1000 new lines of tests covering ProcessSampler, TUI methods, scroll, alerts, run loop
tests/test_swap_clean.py	New tests for short-line skipping, swapoff failure, dry-run target, file-by-file fallback, root check
tests/test_drop_caches.py	New test for OSError propagation from main()
tests/test_sudo_config.py	Import ordering cleanup (ruff-style)
tests/conftest.py	Removed unused io import
pyproject.toml	Added project URLs, coverage config, pytest pythonpath
CHANGELOG.md	Entry for net sort removal
.gitignore	Expanded to standard Python patterns (coverage, packaging, etc.)

[Copilot]

The test asserts "file-by-file" in capsys.readouterr().out, but the key message that proves the file-by-file fallback path was actually triggered is "Not enough RAM to clean all swap at once. Trying file-by-file...", which is printed to stderr (see swap_clean.py line 172). The current assertion happens to pass because the success message "Swap clean completed (file-by-file)." is also on stdout, but it does not actually verify the intended fallback behavior. The assertion should check capsys.readouterr().err to confirm the trigger message is present.