Reviving Praderas (Day 4) — Phase 3: cleaning metadata without drama

Some days are feature launches, others are kitchen cleanup. Phase 3 was the second one: editorial consistency over shiny UI.

What we found during the audit

Before changing content, we audited front matter in content/blog and found:

• 13 posts missing Tags (or using lowercase tags),
• one outlier date format (2023-08-14 9:00),
• canonical taxonomy defined but not automatically enforced.

The content was valuable, but hidden metadata debt would eventually hurt filters, SEO, and category navigation.

What we implemented

1. Completed missing tags on the 13 pending posts using existing taxonomy.
2. Normalized key casing (tags -> Tags) where needed.
3. Standardized date formatting for the outlier case.
4. Added audit script scripts/frontmatter_audit.py to validate:
   • required fields,
   • date format,
   • tags outside canonical set.
5. Documented editorial template in .agents/post-template.md.

Why this matters

Inconsistent metadata often "works" until it quietly fails: partial filters, weak related-post blocks, and blind editorial choices.

This pass improved baseline reliability: fewer surprises, more predictable behavior.

Day 4 closing

Phase 3 is not flashy, but it prevents compound pain. No new screen today; we shipped content discipline.