Skip to content

Growth engineering

Fontana growth engineering turns each MiniSpec and MegaSpec delivery into searchable institutional knowledge. You capture durable fixes and conventions in docs/solutions/; agents consult that store before planning and assurance so the next task starts closer to done.

  • docs/solutions/ - YAML-frontmatter learnings (bug track and knowledge track) organised by category.
  • learnings-researcher - Grep-first search of the solutions store during review, MegaSpec impact/planning, and assurance.
  • /growth-capture - Interactive, headless, or log capture of a solved problem. Every /🫡 log / LOGTHIS runs the mandatory Growth phase: introspect the session, then write docs/solutions/ and focused .cursor/rules so the same class of bug does not recur.
  • Tiered review personas - Always-on validator, reviewer, and learnings search; conditional Convex public-surface and adversarial reviewers when path triggers match.
  1. Gather and implement with MiniSpec (/minispec) or MegaSpec (/megaspec) as usual.
  2. During review and assurance, agents pull prior learnings from docs/solutions/ when relevant.
  3. After assurance and tidy-up, /🫡 log / LOGTHIS always runs the Growth phase (introspect bugs and conventions; write solutions and Cursor rules). Agents record No growth delta only when introspection finds nothing durable.
  • Spec: docs/specs/feature-growth-minispec-megaspec.md (in-repo)
  • Operator guide: docs/solutions/README.md (in-repo)
  • Skill: .cursor/skills/growth/SKILL.md (in-repo)
  • Review persona SSOT: .cursor/skills/megaspec/references/review-personas.md (in-repo)