harness-evolver 4.2.1 → 4.2.2

package/.claude-plugin/plugin.json

@@ -1,7 +1,7 @@
 {
   "name": "harness-evolver",
   "description": "LangSmith-native autonomous agent optimization — evolves LLM agent code using multi-agent proposers, LangSmith experiments, and git worktrees",
-  "version": "4.2.1",
+  "version": "4.2.2",
   "author": {
     "name": "Raphael Valdetaro"
   },

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "harness-evolver",
-  "version": "4.2.1",
+  "version": "4.2.2",
   "description": "LangSmith-native autonomous agent optimization for Claude Code",
   "author": "Raphael Valdetaro",
   "license": "MIT",

package/tools/adversarial_inject.py

@@ -59,7 +59,7 @@ def detect_memorization(client, experiment_name, dataset_name):
     """Check if agent outputs are suspiciously similar to reference outputs."""
     suspicious = []
     try:
-        runs = list(client.list_runs(project_name=experiment_name, is_root=True, limit=200))
+        runs = list(client.list_runs(project_name=experiment_name, is_root=True, limit=100))
         examples = {str(e.id): e for e in client.list_examples(dataset_name=dataset_name, limit=500)}
 
         for run in runs:

package/tools/dataset_health.py

@@ -68,7 +68,7 @@ def check_difficulty(client, config):
         return None
 
     try:
-        runs = list(client.list_runs(project_name=best_exp, is_root=True, limit=200))
+        runs = list(client.list_runs(project_name=best_exp, is_root=True, limit=100))
         if not runs:
             return None
 
@@ -129,7 +129,7 @@ def check_dead_examples(client, config):
 
     for exp_name in recent_exps:
         try:
-            runs = list(client.list_runs(project_name=exp_name, is_root=True, limit=200))
+            runs = list(client.list_runs(project_name=exp_name, is_root=True, limit=100))
             all_run_ids = [run.id for run in runs]
             if not all_run_ids:
                 continue

package/tools/read_results.py

@@ -68,7 +68,7 @@ def read_experiment(client, experiment_name):
         runs = list(client.list_runs(
             project_name=experiment_name,
             is_root=True,
-            limit=200,
+            limit=100,
         ))
 
         if not runs:

package/tools/regression_tracker.py

@@ -60,7 +60,7 @@ def get_per_example_scores(client, experiment_name):
     """Get per-example scores from an experiment."""
     scores = {}
     try:
-        runs = list(client.list_runs(project_name=experiment_name, is_root=True, limit=200))
+        runs = list(client.list_runs(project_name=experiment_name, is_root=True, limit=100))
         all_run_ids = [run.id for run in runs]
         all_feedbacks = list(client.list_feedback(run_ids=all_run_ids))
         fb_map = {}

package/tools/trace_insights.py

@@ -332,7 +332,7 @@ def fetch_scores_from_experiment(experiment_name):
         runs = list(client.list_runs(
             project_name=experiment_name,
             is_root=True,
-            limit=200,
+            limit=100,
         ))
 
         all_run_ids = [run.id for run in runs]