From 6aab9bcfa8363cdc4b6e206ed619806710d45b0e Mon Sep 17 00:00:00 2001
From: Eirik Botten Nicolaysen <eirik@ecodeco.no>
Date: Fri, 10 Jul 2026 13:31:44 +0200
Subject: [PATCH] docs(extensions): add Inspect Judge Drift to the extensions
 gallery

---
 docs/extensions/extensions.yml | 6 ++++++
 1 file changed, 6 insertions(+)

diff --git a/docs/extensions/extensions.yml b/docs/extensions/extensions.yml
index d2994dd51c..902433e3a1 100644
--- a/docs/extensions/extensions.yml
+++ b/docs/extensions/extensions.yml
@@ -82,6 +82,12 @@
   author: "[Eirik Botten Nicolaysen](https://github.com/avalyset)"
   categories: ["Analysis"]
 
+- name: "[Inspect Judge Drift](https://github.com/avalyset/inspect-judge-drift)"
+  description: |
+    Judge-drift measurement — re-grades a stored eval log with two graders over the same samples, so the judge model is the only variable. Grade-parse failures are reported as unscored, never fabricated into verdicts.
+  author: "[Eirik Botten Nicolaysen](https://github.com/avalyset)"
+  categories: ["Analysis"]
+
 - name: "[Inspect Flow](https://meridianlabs-ai.github.io/inspect_flow/)"
   description: |
     Workflow orchestration for reproducibly running evals at scale.