From 742dbe4d41e8aa6b691aea8eb93de9c3017e5990 Mon Sep 17 00:00:00 2001
From: wassname <1103714+wassname@users.noreply.github.com>
Date: Thu, 27 Aug 2026 09:48:21 +0800
Subject: [PATCH] keep the bench numbers out of the skill, point at the README
 table

---
 SKILL.md | 5 ++---
 1 file changed, 2 insertions(+), 3 deletions(-)

diff --git a/SKILL.md b/SKILL.md
index e35f2e4..ac84f29 100644
--- a/SKILL.md
+++ b/SKILL.md
@@ -8,9 +8,8 @@ In an attempt to upskill the machine learning debugging on AI coding assistants
 Be diligent. Work the problem in full before you write. State the decisive point early, then give
 the derivation, the mechanism, or the log line behind it, so the reader can check it and not just
 take it. Show the work, not only the conclusion.
-<!-- CLAUDE: this is at the top because it is the part we measured. A 636 byte document saying
-only this scored +0.66 on wassname-ml-bench v102 (grok-4.6, 12 answers), above bare at +0.44 and
-above this whole skill at +0.35. Do not bury it. -->
+<!-- CLAUDE: top of the file because it is the one part with measured uplift. See the results
+table in README.md. Do not bury it. -->
 
 Then do the exercises for your situation and show the result in your reply. Each
 exercise is marked (small) or (large). A small one takes less than a paragraph to answer; a large