Skip to content

Tier 2 hooks: decompose ~24 non-deterministic guidelines via T1-T7 techniques #11

@lapc506

Description

@lapc506

Goal

Convert remaining non-deterministic dev-discipline guidelines into manifest-driven rules using the T1-T7 decomposition techniques documented in hooks/rules/README.md.

Estimated scope

~24 rules, decomposed via:

  • T1 Narrative → structural assertion (e.g., Slack alerts grouped by project → milestone → assignee)
  • T2 Scope boundary → path-based block (e.g., owner-specific paths)
  • T3 Behavioral mandate → Stop-event reminder (e.g., "don't ask 'should I continue?'")
  • T4 Preference → linter on alternative (e.g., warn on "refactor" in PR titles)
  • T5 Fall-through judgment → explicit gate (e.g., require blocking_reason: when setting blockedBy)
  • T6 Context-sensitive → structural proxy (e.g., detect alert-shaped Slack messages, filter posthog/dashboard)
  • T7 Multi-part → AND-chain of detectors (e.g., owner path + content trigger + exception)

Why follow-up not Tier 1

Tier 2 rules have higher false-positive risk and benefit from Tier 1 field-testing first. Adding them directly to v1.5.0 would balloon the PR diff and dilute review focus.

Approach

  1. Pick one Tier 2 candidate (e.g., linear-blocking-reason via T5).
  2. Decompose into a single rules.yaml row with full tests: array.
  3. Run npm run build-rules && npm run test-hooks.
  4. Open separate small PR (1-3 rules at a time) for incremental review.

Related

Depends on Tier 1 (issue #10) being merged first.

Metadata

Metadata

Assignees

No one assigned

    Labels

    No labels
    No labels

    Type

    No type

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions