feat(rul,cml): durability findings — config lost after compaction (v5.5.0 / A)

Per the official "what survives compaction" model (context-window.md):
only the project-root CLAUDE.md (+ unscoped rules) is re-injected after a
context compaction. Two additive, structural findings (low severity):

- RUL: a large (>50-line) PATH-SCOPED rule — reloads only on a matching
  file read, and is not re-injected after compaction, so must-hold rules
  can silently drop mid-session.
- CML: a NESTED (subdirectory) CLAUDE.md — not re-injected after
  compaction (only the project root is).

Additive (no new scanner, count stays 13); one humanizer entry each;
hermetic temp-fixture tests (positive + negative). Suite 957 -> 961,
SC-5 byte-stable, self-audit A/A.

Known limitation (pre-existing, broader than A): the lightweight
frontmatter parser reads inline `paths:` but not YAML block sequences,
so block-sequence-scoped rules are still seen as unscoped. Deferred.

Part of v5.5.0 "steering-model I". Foundation (active-config-reader
enumeration) deferred to v5.6 with B.

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01Ter3E2JSi1Khgmuf2kady8
This commit is contained in:
Kjell Tore Guttormsen 2026-06-20 11:20:04 +02:00
commit f3aadb5183
6 changed files with 149 additions and 5 deletions

View file

@ -12,7 +12,7 @@
![Commands](https://img.shields.io/badge/commands-18-green)
![Agents](https://img.shields.io/badge/agents-6-orange)
![Hooks](https://img.shields.io/badge/hooks-4-red)
![Tests](https://img.shields.io/badge/tests-957+-brightgreen)
![Tests](https://img.shields.io/badge/tests-961+-brightgreen)
![License](https://img.shields.io/badge/license-MIT-lightgrey)
A Claude Code plugin that checks configuration health, suggests context-aware improvements, and auto-fixes issues — `CLAUDE.md`, `settings.json`, hooks, rules, MCP servers, `@imports`, and plugins. 13 deterministic scanners across 10 quality areas, context-aware feature recommendations, auto-fix with backup/rollback, a prompt-cache-aware Token Hotspots scanner with optional API-calibrated `--accurate-tokens` mode, plus cache-prefix stability, dead-tool, and cross-plugin collision detection. Zero external dependencies.