From 0f7b565bb042b942759156dd9bdf63a6b0cb1e9e Mon Sep 17 00:00:00 2001
From: Bastien Chanot <git@bchanot.fr>
Date: Thu, 30 Jul 2026 12:58:06 +0200
Subject: [PATCH] feat(global): recalibrate instruction layer for Claude 5
 family
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

- Delegation block: model-neutral when-guidance replaces the Opus 4.8
  under-delegation counter (LRN-030 trait inverted on Opus 5; Claude
  Code injects its own anti-delegation prompt there, #80988). Gates
  carve-out keeps verifier/security/challenge dispatch mandatory.
- Drop the 'staff engineer' self-check bar (Opus 5 over-verification
  trigger per official migration guide); honest-reporting steps stay.
- Deviations bullet: finish-whole-task clause (Opus 5 scope-expansion
  counter), scoped so 'gone wrong → STOP' still wins.
- Written-deliverable length rule (Opus 5 writes ~30-40% longer).
---
 CLAUDE.global.md | 16 ++++++++++------
 1 file changed, 10 insertions(+), 6 deletions(-)

diff --git a/CLAUDE.global.md b/CLAUDE.global.md
index 3aa4c2d..3ab8178 100644
--- a/CLAUDE.global.md
+++ b/CLAUDE.global.md
@@ -22,6 +22,8 @@ Apply unless repo-specific instructions override.
 - Document intent, not mechanics. Use project doc style (docstring, JSDoc…).
 - Explicit, consistent, meaningful names. Straight control flow,
   no hidden side effects.
+- Written deliverables (docs, reports, .md): length matched to what
+  the task needs — no filler sections, no boilerplate summaries.
 
 ## Refactoring
 - Priority: safety → readability → consistency.
@@ -40,11 +42,12 @@ Apply unless repo-specific instructions override.
 - Confirm before implementing only when real trade-offs exist (multiple
   valid approaches, breaking change, destructive action) — else proceed.
 - Minimal changes unless broader refactor requested. State trade-offs.
-- Sub-agents keep main context clean — one task per sub-agent.
-  More compute on hard problems. Task fans out across independent
-  items (many files, parallel searches, multi-point checks) → delegate
-  to sub-agents, don't iterate serially. Default to delegation for
-  multi-file exploration. Counters model tendency to under-delegate.
+- Sub-agents: one task per sub-agent, main context stays clean.
+  Delegate genuinely independent, sizeable tracks (wide multi-file
+  exploration, parallel audits) — not work doable in a few tool
+  calls. Skill-mandated gates (fresh verifier/security/challenge)
+  always dispatch as written. Don't redo delegated work by hand —
+  failed gates re-dispatch fresh executors instead.
 - One question upfront if needed — don't interrupt mid-task.
   *Exception: skill-mandated gates and checkpoints (orchestrator
   validation gates, approval gates, darwin checkpoints) always fire.*
@@ -53,6 +56,8 @@ Apply unless repo-specific instructions override.
 - Something goes wrong → STOP, re-plan. Never push through.
 - Deviations: minor or clearly justified → do, explain after.
   Significant or shaky justification → ask before deviating.
+  Finish the whole task: blocked on an independent sub-part → do
+  the rest, state what's missing. Gone WRONG → still STOP, re-plan.
 - Root causes only. No temp fixes. Never assume — verify paths, APIs,
   variables before use.
 
@@ -77,7 +82,6 @@ Apply unless repo-specific instructions override.
 2. Report what verified, what not.
 3. List remaining risks, surviving deviations.
 4. Don't mark complete without proof it works.
-   Bar: "would staff engineer approve?"
 5. Correction or notable event → capitalize to right registry
    (see "Memory registries").