feat(tests): add CI smoke tests for local-execute examples (#237)

[tusharjamunkar]

Description

Fixes #237.

Summary

Adds automated CI smoke testing for local-execute offline demo scripts under examples/. This provides a regression safety net catching import breaks, SkillLoader dispatch regressions, and manifest errors without requiring live API keys or making external network requests.

Key Changes

• tests/test_examples_smoke.py:
  • Parameterized smoke tests covering 8 local-execute offline demo scripts:
    1. mental_coach_demo.py
    2. prompt_injection_firewall_demo.py
    3. prompt_compression_demo.py
    4. deceptive_ui_guard_demo.py
    5. token_limiter_loop.py
    6. uk_companies_house_handler_demo.py
    7. gmail_handler_demo.py
    8. bg_remover_demo.py
  • Added parity test test_every_example_script_has_smoke_or_skip_classification ensuring all scripts in examples/ are categorized with explicit skip rationales or smoke test coverage.
• docs/TESTING.md: Updated status table and four-layer test architecture documentation to document the CI example smoke layer and clarify the skip policy for live model provider loops.
• CHANGELOG.md: Added user-facing entry under [Unreleased].

Verification

• pytest tests/test_examples_smoke.py -v: 9 passed in 11.5s.
• pytest tests/test_registry_docs.py: passed cleanly.
• flake8 tests/test_examples_smoke.py: 0 errors.

[rosspeili]

Nice work @tusharjamunkar, this matches #237 well, and the parity test is a good guard.

Before merge, please:

1. Remove the UTF-8 BOM at the start of tests/test_examples_smoke.py
2. Add a brief note in examples/README.md that local-execute scripts are smoke-tested in CI (link docs/TESTING.md)
3. Optional: bg_remover_demo.py currently passes smoke without calling execute() (missing sample_input.png), either add a tiny fixture or tighten expected markers

pii_guardrail_flow.py skip is fine for v1, we can revisit with a mocked Ollama path later if you wanna open a follow up.

LGTM once the above are fized, thanks! <3

[tusharjamunkar]

Thanks for the review, @rosspeili! All 3 items are addressed:

1. Removed the UTF-8 BOM from tests/test_examples_smoke.py.
2. Added note in examples/README.md referencing CI smoke tests and linking docs/TESTING.md.
3. Tightened expected output markers for bg_remover_demo.py.

All 18 tests and lint checks are passing cleanly!

[tusharjamunkar]

Also applied �lack formatting to ests/test_examples_smoke.py to satisfy CI style checks cleanly.

[rosspeili]

LGTM @tusharjamunkar, merged. <3 Thanks for this.

[tusharjamunkar]

Thanks so much @rosspeili! Glad I could contribute. <3

[rosspeili]

You can pick more good first issues or if you're specifically interested in tests, I would check issues around CLI, Examples, and Testing.

Welcome @tusharjamunkar 🙏