Periodic Review Status

This file tracks when the next periodic repository reviews should be offered. It is current-state scheduling, not a historical log. Git tracks history.

Reviews should be offered at a natural idle point, not during active implementation. The AI may choose the review approach, but should keep findings concise, actionable, and non-duplicative.

AI Architecture Review

• Next prompt after: 2026-05-17
• Cadence: 30 days
• Scope: Review whether the repository is organized so an AI agent can quickly find the right context, understand the system accurately, and avoid unnecessary broad exploration or excessive context use.
• Pay special attention to:
  • Duplicate, overly wordy, stale, or conflicting documentation.
  • Overlap between docs that makes ownership unclear.
  • Missing entry points, unclear architecture boundaries, or misleading file names.
  • Guidance that is too prescriptive, brittle, or likely to block useful AI judgment.
• Output:
  • High-impact findings.
  • Suggested simplifications.
  • Any unclear tradeoffs or questions for the developer team.
  • One suggestion to improve the review process itself.

Repo Guideline Compliance Review

• Next prompt after: 2026-05-17
• Cadence: 30 days
• Scope: Review whether the repository follows its own documented rules, including prototype constraints, README patterns, and agent guidance.
• Pay special attention to:
  • Rules that are duplicated, inconsistent, obsolete, or hard to apply.
  • Docs that disagree with actual code structure or workflows.
  • Places where guidance is too verbose or scattered across multiple files.
  • Missing documentation for conventions that are clearly used in practice.
• Output:
  • Compliance gaps.
  • Suggested doc or code changes.
  • Any questions for the developer team.
  • One suggestion to improve the review process itself.

Code Quality Analysis

• Next prompt after: 2026-05-17
• Cadence: 30 days
• Scope: Run automated code quality analysis and address safe, obvious issues.
• Command: npx fallow --format json
• If the output is unclear, consult fallow.tools.
• Fix safe issues when confidence is high.
• Consider: fallow fix --yes --format json
• Log unresolved questions for the developer team below.

Developer Questions Log

• None currently.