fix(cli): recover classic CLI output after resize

[helix4u]

What does this PR do?

Fixes classic CLI redraw behavior after terminal resize. The prompt_toolkit footer/status chrome was being stamped into normal terminal scrollback during resize storms, and Ctrl+L or /redraw could leave the screen blank because previous output was not rebuildable.

This change keeps a bounded Hermes-owned output history for the classic CLI, debounces resize recovery, suppresses status/footer invalidates while resize recovery is pending, and rebuilds scrollback from that history after the terminal size settles. The resumed-history panel is stored as a structured renderable history entry so it is re-rendered at the current terminal width instead of replayed as stale terminal bytes.

Related Issue

Related to #19280.

Type of Change

• 🐛 Bug fix (non-breaking change that fixes an issue)
• ✨ New feature (non-breaking change that adds functionality)
• 🔒 Security fix
• 📝 Documentation update
• ✅ Tests (adding or improving test coverage)
• ♻️ Refactor (no behavior change)
• 🎯 New skill (bundled or hub)

Changes Made

• cli.py: add bounded replay history for classic CLI output, including structured renderable entries for resize-aware resume panels.
• cli.py: debounce resize recovery and suppress footer/status invalidates while resize recovery is pending.
• cli.py: rebuild terminal scrollback from Hermes-owned history after resize, then repaint the live prompt/footer once at the settled size.
• hermes_cli/config.py: add display.persistent_output and display.persistent_output_max_lines defaults for the replay buffer.
• tests/cli/test_cli_force_redraw.py: cover redraw ordering, resize debounce behavior, and invalidate suppression during resize recovery.
• tests/cli/test_cprint_bg_thread.py: cover output-history recording, replay, suppression, and structured render entries.
• tests/cli/test_resume_display.py: cover storing the resumed-history panel as a resize-aware replay entry.

How to Test

1. Start the classic CLI with a resumed session that displays previous conversation history.
2. Resize the terminal narrower and wider, including while the status timer is updating.
3. Confirm the footer/status area is replaced rather than duplicated, and the resumed-history panel is rebuilt at the current width instead of wrapping stale borders.
4. Run the targeted tests listed below.

Targeted tests run locally:

• HERMES_TEST_WORKERS=1 scripts/run_tests.sh tests/cli/test_resume_display.py tests/cli/test_cli_force_redraw.py tests/cli/test_cprint_bg_thread.py tests/cli/test_cli_init.py tests/cli/test_cli_status_bar.py -q → 115 passed
• HERMES_TEST_WORKERS=1 scripts/run_tests.sh tests/hermes_cli/test_config_drift.py -q → 1 passed

Full suite run after opening draft PR:

• scripts/run_tests.sh → failed: 49 failed, 19874 passed, 51 skipped, 230 warnings in 614.31s
• Failures were outside the modified CLI resize/resume tests and covered existing gateway/provider/browser/delegation/update/TUI areas, including tests/gateway/test_approve_deny_commands.py, tests/cron/test_cron_script.py, tests/agent/test_auxiliary_client.py, tests/agent/test_bedrock_1m_context.py, tests/gateway/test_dingtalk.py, tests/hermes_cli/test_update_gateway_restart.py, tests/tools/test_browser_chromium_check.py, tests/tools/test_delegate.py, and tests/tui_gateway/test_goal_command.py.

Checklist

Code

• I've read the Contributing Guide
• My commit messages follow Conventional Commits (fix(scope):, feat(scope):, etc.)
• I searched for existing PRs to make sure this isn't a duplicate
• My PR contains only changes related to this fix/feature (no unrelated commits)
• I've run pytest tests/ -q and all tests pass
• I've added tests for my changes (required for bug fixes, strongly encouraged for features)
• I've tested on my platform: Windows Terminal / Ubuntu under WSL

Documentation & Housekeeping

• I've updated relevant documentation (README, docs/, docstrings) — or N/A
• I've updated cli-config.yaml.example if I added/changed config keys — or N/A
• I've updated CONTRIBUTING.md or AGENTS.md if I changed architecture or workflows — or N/A
• I've considered cross-platform impact (Windows, macOS) per the compatibility guide — or N/A
• I've updated tool descriptions/schemas if I changed tool behavior — or N/A

Screenshots / Logs

Manual reproduction involved Windows Terminal resizing of the classic CLI with a resumed session and live status/footer updates. Targeted and full-suite test output is summarized above.