v0.6.0

v0.6.0 — Rendered-site overhaul + ftg render

A focused release on the report itself: the rendered site got a substantial visual overhaul, and you can now rebuild it from cached findings without re-running analysis.

Highlights

• 🎨 Site visual overhaul. New layout, sharper typography, finding cards with priority stripes, and a tightened nav. The same theme later becomes the visual reference for the PDF export in v0.14.0.
• 🛠️ ftg render subcommand. Rebuild the Hugo report site from a cached analyze run — useful when you tweak templates/CSS and want the new look without burning LLM calls.
• 🚫 Filter out marketing pages and blog posts. Spider classification now excludes /blog/, /customers/, /team/, and similar non-docs sections so drift and screenshot findings don't fire on press releases.
• 🩹 Tolerant schema validation. Structured-output validation now ignores extra fields instead of failing the whole call, which sidesteps a class of provider-side schema drift.

What's Changed

• fix(analyzer): tolerate extra fields in JSON schema validation by @britt in #63
• feat(site): visual overhaul of the rendered site by @britt in #64
• feat(cli): add ftg render to rebuild the site from cached findings by @britt in #65
• feat(analyzer): exclude marketing pages and blog posts from docs by @britt in #66

Full Changelog: v0.5.1...v0.6.0