Convert quality criteria from questions to statements

[nhorton]

Summary

This PR converts all quality criteria throughout the codebase from question format to statement format, describing the expected state after a step completes rather than asking whether that state was achieved.

Key Changes

Documentation and Examples

• Updated job.yml files and templates to use statements instead of questions in quality_criteria

  • Example: "Does the job break out across logical steps..." → "The job breaks out across logical steps..."
  • Applied consistently across all quality criteria in standard jobs and examples

• Updated instructional documentation in define.md, implement.md, iterate.md, learn.md, and fix_jobs.md to reflect the new statement-based format with examples

• Updated best practices documentation in research_report_job_best_practices.md with statement-based criteria

Code Changes

• quality_gate.py:

  • Added optional notes parameter to _build_payload() method to support author notes in review payloads
  • Updated build_review_instructions_file() to handle author notes separately from outputs
  • Increased base timeout from 120 seconds to 240 seconds (4 minutes) for review operations
  • Updated timeout calculation examples in docstring to reflect new base timeout

• test_quality_gate.py:

  • Updated all timeout tests to reflect new 240-second base timeout
  • Updated test assertions and comments to match new timeout calculations

Templates

• Updated job.yml.template to clarify that quality criteria should be statements of expected state, not questions

Implementation Details

• The change maintains backward compatibility in the code while updating all user-facing documentation and examples
• The timeout increase provides more time for reviewers to evaluate complex outputs
• Author notes are now properly separated from output file contents in review payloads

https://claude.ai/code/session_01Jn2nuSQ3rUdhcr29s13K9T