[Grok] Best ACM dao-shu guidance (#10 + #11 layers)

[KorenKrita]

Summary

Best always-on ACM guidance: #10 Grok perfect spine + cherry-pick of #11’s only unique win (judgment-layers map).

From #10 (kept)

• Corrective Failure shapes (what to do next)
• Fold targets by named boundary only; cold start on rebase candidate compare
• Split skill disclosure: ambiguous bases inline under rebase; end list for archive/exception only
• Rebase candidate compare sentence vs proximity

From #11 (cherry-picked)

• How judgment layers fit together (smells → moments → criteria → failure shapes)
• Fuller causal opening (Process detail accumulates + ACM keeps that growth recoverable)
• without rereading archived summaries
• Skill router mentions judgment layers

Explicitly not taken from #11

• Dual ambiguous-base skill pointers
• Cold start on ordinary fold target line
• Thin diagnostic-only failure shapes

CORE length lands ~9300 (soft <10000). Recommend superseding #2–#11 after review.

Test plan

• bun run verify:acm
• Live: read layers map once → smell → moment → criteria → travel; failure shape suggests correction

[KorenKrita]

Superseded by the compression-as-intelligence redesign on main, which replaced the layered dao-shu guidance drafts with the CORE.md + TOOL-CONTRACTS.md canonical split.