test: add extensible browser evaluation harness

2026-09-10 13:01:23 +08:00 · 2026-08-28 12:16:39 +08:00
parent 7aaa5a2d01
commit 9ad33801c7
46 changed files with 3554 additions and 0 deletions
@@ -0,0 +1,17 @@
+{
+  "agents": {
+    "dsh-local": {
+      "variant": "granular-28",
+      "command": "dsh",
+      "args": ["--profile", "headless", "{prompt}"],
+      "cwd": "../..",
+      "timeoutMs": 300000,
+      "metadata": {
+        "toolCount": 28
+      },
+      "metrics": {
+        "toolCallPattern": "browser[_-][a-z_-]+"
+      }
+    }
+  }
+}