Skip to content

Mine the MS 5-category / 45-rule taxonomy for falsifiable rules we lack [R2 from #340] #429

Description

@russmiles

Follow-up R2 from the microsoft/AI-Engineering-Coach exploration (docs/superpowers/explorations/2026-06-16-ai-engineering-coach-findings.md, issue #340).

Idea

The MS project ships ~45 anti-pattern rules across 5 practice-score categories (Prompt Quality, Session Hygiene, Code Review, Tool Mastery, Context Management). Treat that taxonomy as a cross-check checklist: walk it against our GC catalogue (skills/garbage-collection/references/gc-catalogue.md), governance anti-pattern gallery, and /assess evidence signals, and adopt — in our own falsifiable-constraint idiom — any genuinely-missing checks.

Scope

  • Read the actual rule files in the MS repo (the README does not document the rule schema).
  • For each MS rule, classify: already-covered / additive / not-applicable-to-our-substrate (many are session-log-based → defer to R1).
  • Author any adopted checks as new GC rules or assessment signals.

Attribution

Ideas are not copyrightable. If any rule text is copied near-verbatim, add an MIT NOTICE in the same PR (see the findings doc's attribution section). Prefer re-deriving to copying.

Activity

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Metadata

Metadata

Assignees

No one assigned

    Labels

    enhancementNew feature or request

    Type

    No type

    Projects

    No projects

      Milestone

      No milestone

      Relationships

      None yet

      Development

      No branches or pull requests

      Issue actions