docs: document pipeline artifact contract and runtime outputs

[adidev001]

Summary

This PR updates docs/pipeline.md to make the pipeline artifact layout a bit clearer and closer to the outputs contributors actually work with in practice.

What changed

• added clean_markdown/ to the artifact layout
• documented runtime/resume state like .processing_state.pkl
• documented debug folders such as problematic_files/ and timeout_files/
• clarified which artifacts are core pipeline outputs versus runtime/debug outputs
• added a reminder to keep docs/pipeline.md and tests/test_pipeline_smoke.py aligned when artifact paths change

Why

While going through the docs and smoke tests during local setup, I noticed that some contributor-relevant runtime artifacts were reflected in tests and contribution notes, but not as clearly in the main pipeline layout documentation.

This change is meant to make that stage-level contract easier to follow for future contributors, especially when debugging pipeline runs or updating output paths.

[fffoivos]

Thanks for the improvemetns @adidev001

[adidev001]

Thanks for the improvemetns @adidev001

Thanks a lot for reviewing and merging this — really appreciate it!