opencode-swarm-plugin 0.39.1 → 0.42.0

package/.opencode/eval-history.jsonl

@@ -10,3 +10,318 @@
 {"timestamp":"2025-12-25T05:11:18.469Z","eval_name":"coordinator-behavior","score":0.85,"run_count":3}
 {"timestamp":"2025-12-25T05:11:18.469Z","eval_name":"coordinator-session","score":0.85,"run_count":3}
 {"timestamp":"2025-12-25T05:11:18.469Z","eval_name":"swarm-decomposition","score":0.85,"run_count":3}
+{"timestamp":"2025-12-25T16:30:42.957Z","eval_name":"Coordinator Discipline - Synthetic Fixtures","score":0.6354444444444443,"run_count":1}
+{"timestamp":"2025-12-25T16:30:42.957Z","eval_name":"Coordinator Discipline - Real Sessions","score":0,"run_count":1}
+{"timestamp":"2025-12-25T16:30:42.957Z","eval_name":"Coordinator Discipline - Perfect vs Bad","score":0.5416666666666666,"run_count":1}
+{"timestamp":"2025-12-25T16:30:42.957Z","eval_name":"Swarm Decomposition Quality","score":0.7213888888888889,"run_count":1}
+{"timestamp":"2025-12-25T16:30:42.957Z","eval_name":"Decomposition Edge Cases","score":0.775,"run_count":1}
+{"timestamp":"2025-12-25T16:30:42.957Z","eval_name":"Compaction Hook Coordinator Resumption","score":0.95,"run_count":1}
+{"timestamp":"2025-12-25T16:30:42.957Z","eval_name":"Epic ID Specificity","score":0.5,"run_count":1}
+{"timestamp":"2025-12-25T16:30:42.958Z","eval_name":"No False Positives","score":1,"run_count":1}
+{"timestamp":"2025-12-25T16:30:42.958Z","eval_name":"Example: Basic scorer test","score":1,"run_count":1}
+{"timestamp":"2025-12-25T16:30:42.958Z","eval_name":"Perfect Prompt Scores 100%","score":1,"run_count":1}
+{"timestamp":"2025-12-25T16:30:42.958Z","eval_name":"Placeholder Detection","score":0,"run_count":1}
+{"timestamp":"2025-12-25T16:30:42.958Z","eval_name":"Generic Instructions Fail","score":0,"run_count":1}
+{"timestamp":"2025-12-25T16:30:42.958Z","eval_name":"First Tool Discipline","score":0,"run_count":1}
+{"timestamp":"2025-12-25T16:30:42.958Z","eval_name":"Compaction Prompt Quality","score":0.6342857142857142,"run_count":1}
+{"timestamp":"2025-12-25T16:30:42.958Z","eval_name":"Coordinator Resists Direct Implementation","score":1,"run_count":1}
+{"timestamp":"2025-12-25T16:30:42.958Z","eval_name":"Coordinator Behavior After Compaction","score":0.8645833333333333,"run_count":1}
+{"timestamp":"2025-12-25T16:30:43.088Z","eval_name":"Example: Basic scorer test","score":1,"run_count":2}
+{"timestamp":"2025-12-25T16:30:43.202Z","eval_name":"Example: Basic scorer test","score":1,"run_count":3}
+{"timestamp":"2025-12-25T16:30:43.316Z","eval_name":"Example: Basic scorer test","score":1,"run_count":4}
+{"timestamp":"2025-12-25T16:31:17.738Z","eval_name":"Coordinator Discipline - Synthetic Fixtures","score":0.6354444444444443,"run_count":2}
+{"timestamp":"2025-12-25T16:31:17.738Z","eval_name":"Coordinator Discipline - Real Sessions","score":0,"run_count":2}
+{"timestamp":"2025-12-25T16:31:17.738Z","eval_name":"Coordinator Discipline - Perfect vs Bad","score":0.5416666666666666,"run_count":2}
+{"timestamp":"2025-12-25T16:31:17.738Z","eval_name":"Swarm Decomposition Quality","score":0.6748148148148146,"run_count":2}
+{"timestamp":"2025-12-25T16:31:17.738Z","eval_name":"Decomposition Edge Cases","score":0.775,"run_count":2}
+{"timestamp":"2025-12-25T16:31:17.738Z","eval_name":"Compaction Hook Coordinator Resumption","score":0.95,"run_count":2}
+{"timestamp":"2025-12-25T16:31:17.738Z","eval_name":"Epic ID Specificity","score":0.5,"run_count":2}
+{"timestamp":"2025-12-25T16:31:17.738Z","eval_name":"No False Positives","score":1,"run_count":2}
+{"timestamp":"2025-12-25T16:31:17.738Z","eval_name":"Example: Basic scorer test","score":1,"run_count":5}
+{"timestamp":"2025-12-25T16:31:17.738Z","eval_name":"Coordinator Behavior After Compaction","score":1,"run_count":2}
+{"timestamp":"2025-12-25T16:31:17.738Z","eval_name":"Coordinator Resists Direct Implementation","score":1,"run_count":2}
+{"timestamp":"2025-12-25T16:31:17.738Z","eval_name":"Perfect Prompt Scores 100%","score":1,"run_count":2}
+{"timestamp":"2025-12-25T16:31:17.738Z","eval_name":"Placeholder Detection","score":0,"run_count":2}
+{"timestamp":"2025-12-25T16:31:17.738Z","eval_name":"Generic Instructions Fail","score":0,"run_count":2}
+{"timestamp":"2025-12-25T16:31:17.738Z","eval_name":"First Tool Discipline","score":0,"run_count":2}
+{"timestamp":"2025-12-25T16:31:17.738Z","eval_name":"Compaction Prompt Quality","score":0.6342857142857142,"run_count":2}
+{"timestamp":"2025-12-25T16:31:17.874Z","eval_name":"Example: Basic scorer test","score":1,"run_count":6}
+{"timestamp":"2025-12-25T16:31:17.995Z","eval_name":"Example: Basic scorer test","score":1,"run_count":7}
+{"timestamp":"2025-12-25T16:31:18.113Z","eval_name":"Example: Basic scorer test","score":1,"run_count":8}
+{"timestamp":"2025-12-25T16:32:49.807Z","eval_name":"Coordinator Discipline - Synthetic Fixtures","score":0.6354444444444443,"run_count":3}
+{"timestamp":"2025-12-25T16:32:49.807Z","eval_name":"Coordinator Discipline - Real Sessions","score":0,"run_count":3}
+{"timestamp":"2025-12-25T16:32:49.807Z","eval_name":"Coordinator Discipline - Perfect vs Bad","score":0.5416666666666666,"run_count":3}
+{"timestamp":"2025-12-25T16:32:49.807Z","eval_name":"Swarm Decomposition Quality","score":0.6988888888888889,"run_count":3}
+{"timestamp":"2025-12-25T16:32:49.807Z","eval_name":"Decomposition Edge Cases","score":0.775,"run_count":3}
+{"timestamp":"2025-12-25T16:32:49.807Z","eval_name":"Compaction Hook Coordinator Resumption","score":0.95,"run_count":3}
+{"timestamp":"2025-12-25T16:32:49.807Z","eval_name":"Epic ID Specificity","score":0.5,"run_count":3}
+{"timestamp":"2025-12-25T16:32:49.807Z","eval_name":"No False Positives","score":1,"run_count":3}
+{"timestamp":"2025-12-25T16:32:49.807Z","eval_name":"Example: Basic scorer test","score":1,"run_count":9}
+{"timestamp":"2025-12-25T16:32:49.807Z","eval_name":"Compaction Prompt Quality","score":0.6342857142857142,"run_count":3}
+{"timestamp":"2025-12-25T16:32:49.807Z","eval_name":"Perfect Prompt Scores 100%","score":1,"run_count":3}
+{"timestamp":"2025-12-25T16:32:49.807Z","eval_name":"Placeholder Detection","score":0,"run_count":3}
+{"timestamp":"2025-12-25T16:32:49.808Z","eval_name":"Generic Instructions Fail","score":0,"run_count":3}
+{"timestamp":"2025-12-25T16:32:49.808Z","eval_name":"First Tool Discipline","score":0,"run_count":3}
+{"timestamp":"2025-12-25T16:32:49.808Z","eval_name":"Coordinator Behavior After Compaction","score":1,"run_count":3}
+{"timestamp":"2025-12-25T16:32:49.808Z","eval_name":"Coordinator Resists Direct Implementation","score":0.9375,"run_count":3}
+{"timestamp":"2025-12-25T16:32:49.943Z","eval_name":"Example: Basic scorer test","score":1,"run_count":10}
+{"timestamp":"2025-12-25T16:32:50.073Z","eval_name":"Example: Basic scorer test","score":1,"run_count":11}
+{"timestamp":"2025-12-25T16:32:50.199Z","eval_name":"Example: Basic scorer test","score":1,"run_count":12}
+{"timestamp":"2025-12-25T16:35:31.083Z","eval_name":"Coordinator Discipline - Synthetic Fixtures","score":0.6354444444444443,"run_count":4}
+{"timestamp":"2025-12-25T16:35:31.083Z","eval_name":"Coordinator Discipline - Real Sessions","score":0,"run_count":4}
+{"timestamp":"2025-12-25T16:35:31.083Z","eval_name":"Coordinator Discipline - Perfect vs Bad","score":0.5416666666666666,"run_count":4}
+{"timestamp":"2025-12-25T16:35:31.083Z","eval_name":"Swarm Decomposition Quality","score":0.6798611111111109,"run_count":4}
+{"timestamp":"2025-12-25T16:35:31.083Z","eval_name":"Decomposition Edge Cases","score":0.775,"run_count":4}
+{"timestamp":"2025-12-25T16:35:31.083Z","eval_name":"Compaction Hook Coordinator Resumption","score":0.95,"run_count":4}
+{"timestamp":"2025-12-25T16:35:31.083Z","eval_name":"Epic ID Specificity","score":0.5,"run_count":4}
+{"timestamp":"2025-12-25T16:35:31.083Z","eval_name":"No False Positives","score":1,"run_count":4}
+{"timestamp":"2025-12-25T16:35:31.083Z","eval_name":"Example: Basic scorer test","score":1,"run_count":13}
+{"timestamp":"2025-12-25T16:35:31.084Z","eval_name":"Compaction Prompt Quality","score":0.6342857142857142,"run_count":4}
+{"timestamp":"2025-12-25T16:35:31.084Z","eval_name":"Perfect Prompt Scores 100%","score":1,"run_count":4}
+{"timestamp":"2025-12-25T16:35:31.084Z","eval_name":"Placeholder Detection","score":0,"run_count":4}
+{"timestamp":"2025-12-25T16:35:31.084Z","eval_name":"Generic Instructions Fail","score":0,"run_count":4}
+{"timestamp":"2025-12-25T16:35:31.084Z","eval_name":"First Tool Discipline","score":0,"run_count":4}
+{"timestamp":"2025-12-25T16:35:31.084Z","eval_name":"Coordinator Resists Direct Implementation","score":0.9375,"run_count":4}
+{"timestamp":"2025-12-25T16:35:31.084Z","eval_name":"Coordinator Behavior After Compaction","score":0.8645833333333333,"run_count":4}
+{"timestamp":"2025-12-25T16:35:31.221Z","eval_name":"Example: Basic scorer test","score":1,"run_count":14}
+{"timestamp":"2025-12-25T16:35:31.329Z","eval_name":"Example: Basic scorer test","score":1,"run_count":15}
+{"timestamp":"2025-12-25T16:35:31.444Z","eval_name":"Example: Basic scorer test","score":1,"run_count":16}
+{"timestamp":"2025-12-25T16:35:31.685Z","eval_name":"Example: Basic scorer test","score":1,"run_count":17}
+{"timestamp":"2025-12-25T16:35:31.843Z","eval_name":"Example: Basic scorer test","score":1,"run_count":18}
+{"timestamp":"2025-12-25T16:35:31.962Z","eval_name":"Example: Basic scorer test","score":1,"run_count":19}
+{"timestamp":"2025-12-25T16:35:32.076Z","eval_name":"Example: Basic scorer test","score":1,"run_count":20}
+{"timestamp":"2025-12-25T16:36:03.596Z","eval_name":"Coordinator Discipline - Synthetic Fixtures","score":0.6354444444444443,"run_count":5}
+{"timestamp":"2025-12-25T16:36:03.596Z","eval_name":"Coordinator Discipline - Real Sessions","score":0,"run_count":5}
+{"timestamp":"2025-12-25T16:36:03.597Z","eval_name":"Coordinator Discipline - Perfect vs Bad","score":0.5416666666666666,"run_count":5}
+{"timestamp":"2025-12-25T16:36:03.597Z","eval_name":"Swarm Decomposition Quality","score":0.6845833333333333,"run_count":5}
+{"timestamp":"2025-12-25T16:36:03.597Z","eval_name":"Decomposition Edge Cases","score":0.775,"run_count":5}
+{"timestamp":"2025-12-25T16:36:03.597Z","eval_name":"Compaction Hook Coordinator Resumption","score":0.95,"run_count":5}
+{"timestamp":"2025-12-25T16:36:03.597Z","eval_name":"Epic ID Specificity","score":0.5,"run_count":5}
+{"timestamp":"2025-12-25T16:36:03.597Z","eval_name":"No False Positives","score":1,"run_count":5}
+{"timestamp":"2025-12-25T16:36:03.597Z","eval_name":"Example: Basic scorer test","score":1,"run_count":21}
+{"timestamp":"2025-12-25T16:36:03.597Z","eval_name":"Compaction Prompt Quality","score":0.6342857142857142,"run_count":5}
+{"timestamp":"2025-12-25T16:36:03.597Z","eval_name":"Perfect Prompt Scores 100%","score":1,"run_count":5}
+{"timestamp":"2025-12-25T16:36:03.597Z","eval_name":"Placeholder Detection","score":0,"run_count":5}
+{"timestamp":"2025-12-25T16:36:03.597Z","eval_name":"Generic Instructions Fail","score":0,"run_count":5}
+{"timestamp":"2025-12-25T16:36:03.597Z","eval_name":"First Tool Discipline","score":0,"run_count":5}
+{"timestamp":"2025-12-25T16:36:03.597Z","eval_name":"Coordinator Behavior After Compaction","score":1,"run_count":5}
+{"timestamp":"2025-12-25T16:36:03.597Z","eval_name":"Coordinator Resists Direct Implementation","score":1,"run_count":5}
+{"timestamp":"2025-12-25T16:36:03.727Z","eval_name":"Example: Basic scorer test","score":1,"run_count":22}
+{"timestamp":"2025-12-25T16:36:03.842Z","eval_name":"Example: Basic scorer test","score":1,"run_count":23}
+{"timestamp":"2025-12-25T16:36:03.951Z","eval_name":"Example: Basic scorer test","score":1,"run_count":24}
+{"timestamp":"2025-12-25T16:36:04.183Z","eval_name":"Example: Basic scorer test","score":1,"run_count":25}
+{"timestamp":"2025-12-25T16:36:04.330Z","eval_name":"Example: Basic scorer test","score":1,"run_count":26}
+{"timestamp":"2025-12-25T16:36:04.445Z","eval_name":"Example: Basic scorer test","score":1,"run_count":27}
+{"timestamp":"2025-12-25T16:36:04.555Z","eval_name":"Example: Basic scorer test","score":1,"run_count":28}
+{"timestamp":"2025-12-25T16:36:32.176Z","eval_name":"Coordinator Discipline - Synthetic Fixtures","score":0.6354444444444443,"run_count":6}
+{"timestamp":"2025-12-25T16:36:32.176Z","eval_name":"Coordinator Discipline - Real Sessions","score":0,"run_count":6}
+{"timestamp":"2025-12-25T16:36:32.176Z","eval_name":"Coordinator Discipline - Perfect vs Bad","score":0.5416666666666666,"run_count":6}
+{"timestamp":"2025-12-25T16:36:32.176Z","eval_name":"Swarm Decomposition Quality","score":0.6852777777777778,"run_count":6}
+{"timestamp":"2025-12-25T16:36:32.176Z","eval_name":"Decomposition Edge Cases","score":0.775,"run_count":6}
+{"timestamp":"2025-12-25T16:36:32.176Z","eval_name":"Epic ID Specificity","score":0.5,"run_count":6}
+{"timestamp":"2025-12-25T16:36:32.176Z","eval_name":"No False Positives","score":1,"run_count":6}
+{"timestamp":"2025-12-25T16:36:32.177Z","eval_name":"Example: Basic scorer test","score":1,"run_count":29}
+{"timestamp":"2025-12-25T16:36:32.177Z","eval_name":"Compaction Hook Coordinator Resumption","score":0.95,"run_count":6}
+{"timestamp":"2025-12-25T16:36:32.177Z","eval_name":"Perfect Prompt Scores 100%","score":1,"run_count":6}
+{"timestamp":"2025-12-25T16:36:32.177Z","eval_name":"Placeholder Detection","score":0,"run_count":6}
+{"timestamp":"2025-12-25T16:36:32.177Z","eval_name":"Generic Instructions Fail","score":0,"run_count":6}
+{"timestamp":"2025-12-25T16:36:32.177Z","eval_name":"First Tool Discipline","score":0,"run_count":6}
+{"timestamp":"2025-12-25T16:36:32.177Z","eval_name":"Coordinator Behavior After Compaction","score":0.978125,"run_count":6}
+{"timestamp":"2025-12-25T16:36:32.177Z","eval_name":"Coordinator Resists Direct Implementation","score":1,"run_count":6}
+{"timestamp":"2025-12-25T16:36:32.177Z","eval_name":"Compaction Prompt Quality","score":0.6342857142857142,"run_count":6}
+{"timestamp":"2025-12-25T16:36:32.305Z","eval_name":"Example: Basic scorer test","score":1,"run_count":30}
+{"timestamp":"2025-12-25T16:36:32.416Z","eval_name":"Example: Basic scorer test","score":1,"run_count":31}
+{"timestamp":"2025-12-25T16:36:32.527Z","eval_name":"Example: Basic scorer test","score":1,"run_count":32}
+{"timestamp":"2025-12-25T16:36:32.755Z","eval_name":"Example: Basic scorer test","score":1,"run_count":33}
+{"timestamp":"2025-12-25T16:36:32.957Z","eval_name":"Example: Basic scorer test","score":1,"run_count":34}
+{"timestamp":"2025-12-25T16:36:33.071Z","eval_name":"Example: Basic scorer test","score":1,"run_count":35}
+{"timestamp":"2025-12-25T16:36:33.180Z","eval_name":"Example: Basic scorer test","score":1,"run_count":36}
+{"timestamp":"2025-12-25T16:38:02.146Z","eval_name":"Coordinator Discipline - Perfect vs Bad","score":0.5416666666666666,"run_count":7}
+{"timestamp":"2025-12-25T16:38:02.147Z","eval_name":"Coordinator Discipline - Synthetic Fixtures","score":0.6354444444444443,"run_count":7}
+{"timestamp":"2025-12-25T16:38:02.147Z","eval_name":"Coordinator Discipline - Real Sessions","score":0,"run_count":7}
+{"timestamp":"2025-12-25T16:38:02.147Z","eval_name":"Swarm Decomposition Quality","score":0.6726388888888888,"run_count":7}
+{"timestamp":"2025-12-25T16:38:02.147Z","eval_name":"Decomposition Edge Cases","score":0.775,"run_count":7}
+{"timestamp":"2025-12-25T16:38:02.147Z","eval_name":"Compaction Hook Coordinator Resumption","score":0.95,"run_count":7}
+{"timestamp":"2025-12-25T16:38:02.147Z","eval_name":"Epic ID Specificity","score":0.5,"run_count":7}
+{"timestamp":"2025-12-25T16:38:02.147Z","eval_name":"No False Positives","score":1,"run_count":7}
+{"timestamp":"2025-12-25T16:38:02.147Z","eval_name":"Example: Basic scorer test","score":1,"run_count":37}
+{"timestamp":"2025-12-25T16:38:02.147Z","eval_name":"Compaction Prompt Quality","score":0.6342857142857142,"run_count":7}
+{"timestamp":"2025-12-25T16:38:02.147Z","eval_name":"Perfect Prompt Scores 100%","score":1,"run_count":7}
+{"timestamp":"2025-12-25T16:38:02.147Z","eval_name":"Placeholder Detection","score":0,"run_count":7}
+{"timestamp":"2025-12-25T16:38:02.147Z","eval_name":"Generic Instructions Fail","score":0,"run_count":7}
+{"timestamp":"2025-12-25T16:38:02.147Z","eval_name":"First Tool Discipline","score":0,"run_count":7}
+{"timestamp":"2025-12-25T16:38:02.148Z","eval_name":"Coordinator Behavior After Compaction","score":0.8645833333333333,"run_count":7}
+{"timestamp":"2025-12-25T16:38:02.148Z","eval_name":"Coordinator Resists Direct Implementation","score":1,"run_count":7}
+{"timestamp":"2025-12-25T16:38:02.276Z","eval_name":"Example: Basic scorer test","score":1,"run_count":38}
+{"timestamp":"2025-12-25T16:38:02.428Z","eval_name":"Example: Basic scorer test","score":1,"run_count":39}
+{"timestamp":"2025-12-25T16:38:02.547Z","eval_name":"Example: Basic scorer test","score":1,"run_count":40}
+{"timestamp":"2025-12-25T16:38:02.782Z","eval_name":"Example: Basic scorer test","score":1,"run_count":41}
+{"timestamp":"2025-12-25T16:38:02.933Z","eval_name":"Example: Basic scorer test","score":1,"run_count":42}
+{"timestamp":"2025-12-25T16:38:03.050Z","eval_name":"Example: Basic scorer test","score":1,"run_count":43}
+{"timestamp":"2025-12-25T16:38:03.165Z","eval_name":"Example: Basic scorer test","score":1,"run_count":44}
+{"timestamp":"2025-12-25T16:38:52.756Z","eval_name":"Coordinator Discipline - Real Sessions","score":0,"run_count":8}
+{"timestamp":"2025-12-25T16:38:52.757Z","eval_name":"Coordinator Discipline - Perfect vs Bad","score":0.5416666666666666,"run_count":8}
+{"timestamp":"2025-12-25T16:38:52.757Z","eval_name":"Coordinator Discipline - Synthetic Fixtures","score":0.6354444444444443,"run_count":8}
+{"timestamp":"2025-12-25T16:38:52.757Z","eval_name":"Swarm Decomposition Quality","score":0.695,"run_count":8}
+{"timestamp":"2025-12-25T16:38:52.757Z","eval_name":"Decomposition Edge Cases","score":0.775,"run_count":8}
+{"timestamp":"2025-12-25T16:38:52.757Z","eval_name":"Compaction Hook Coordinator Resumption","score":0.95,"run_count":8}
+{"timestamp":"2025-12-25T16:38:52.757Z","eval_name":"Epic ID Specificity","score":0.5,"run_count":8}
+{"timestamp":"2025-12-25T16:38:52.757Z","eval_name":"No False Positives","score":1,"run_count":8}
+{"timestamp":"2025-12-25T16:38:52.757Z","eval_name":"Example: Basic scorer test","score":1,"run_count":45}
+{"timestamp":"2025-12-25T16:38:52.757Z","eval_name":"Perfect Prompt Scores 100%","score":1,"run_count":8}
+{"timestamp":"2025-12-25T16:38:52.757Z","eval_name":"Placeholder Detection","score":0,"run_count":8}
+{"timestamp":"2025-12-25T16:38:52.757Z","eval_name":"Generic Instructions Fail","score":0,"run_count":8}
+{"timestamp":"2025-12-25T16:38:52.757Z","eval_name":"First Tool Discipline","score":0,"run_count":8}
+{"timestamp":"2025-12-25T16:38:52.757Z","eval_name":"Compaction Prompt Quality","score":0.6342857142857142,"run_count":8}
+{"timestamp":"2025-12-25T16:38:52.757Z","eval_name":"Coordinator Behavior After Compaction","score":0.9526041666666667,"run_count":8}
+{"timestamp":"2025-12-25T16:38:52.758Z","eval_name":"Coordinator Resists Direct Implementation","score":1,"run_count":8}
+{"timestamp":"2025-12-25T16:38:52.903Z","eval_name":"Example: Basic scorer test","score":1,"run_count":46}
+{"timestamp":"2025-12-25T16:38:53.020Z","eval_name":"Example: Basic scorer test","score":1,"run_count":47}
+{"timestamp":"2025-12-25T16:38:53.136Z","eval_name":"Example: Basic scorer test","score":1,"run_count":48}
+{"timestamp":"2025-12-25T16:38:53.367Z","eval_name":"Example: Basic scorer test","score":1,"run_count":49}
+{"timestamp":"2025-12-25T16:38:53.511Z","eval_name":"Example: Basic scorer test","score":1,"run_count":50}
+{"timestamp":"2025-12-25T16:38:53.624Z","eval_name":"Example: Basic scorer test","score":1,"run_count":51}
+{"timestamp":"2025-12-25T16:38:53.737Z","eval_name":"Example: Basic scorer test","score":1,"run_count":52}
+{"timestamp":"2025-12-25T16:40:39.219Z","eval_name":"Coordinator Discipline - Synthetic Fixtures","score":0.6354444444444443,"run_count":9}
+{"timestamp":"2025-12-25T16:40:39.219Z","eval_name":"Coordinator Discipline - Real Sessions","score":0,"run_count":9}
+{"timestamp":"2025-12-25T16:40:39.219Z","eval_name":"Coordinator Discipline - Perfect vs Bad","score":0.5416666666666666,"run_count":9}
+{"timestamp":"2025-12-25T16:40:39.219Z","eval_name":"Decomposition Edge Cases","score":0.775,"run_count":9}
+{"timestamp":"2025-12-25T16:40:39.220Z","eval_name":"Swarm Decomposition Quality","score":0.7020833333333334,"run_count":9}
+{"timestamp":"2025-12-25T16:40:39.220Z","eval_name":"Epic ID Specificity","score":0.5,"run_count":9}
+{"timestamp":"2025-12-25T16:40:39.220Z","eval_name":"No False Positives","score":1,"run_count":9}
+{"timestamp":"2025-12-25T16:40:39.220Z","eval_name":"Compaction Hook Coordinator Resumption","score":0.95,"run_count":9}
+{"timestamp":"2025-12-25T16:40:39.220Z","eval_name":"Example: Basic scorer test","score":1,"run_count":53}
+{"timestamp":"2025-12-25T16:40:39.220Z","eval_name":"Coordinator Behavior After Compaction","score":0.7291666666666666,"run_count":9}
+{"timestamp":"2025-12-25T16:40:39.220Z","eval_name":"Coordinator Resists Direct Implementation","score":1,"run_count":9}
+{"timestamp":"2025-12-25T16:40:39.220Z","eval_name":"Compaction Prompt Quality","score":0.6342857142857142,"run_count":9}
+{"timestamp":"2025-12-25T16:40:39.220Z","eval_name":"Perfect Prompt Scores 100%","score":1,"run_count":9}
+{"timestamp":"2025-12-25T16:40:39.220Z","eval_name":"Placeholder Detection","score":0,"run_count":9}
+{"timestamp":"2025-12-25T16:40:39.220Z","eval_name":"Generic Instructions Fail","score":0,"run_count":9}
+{"timestamp":"2025-12-25T16:40:39.220Z","eval_name":"First Tool Discipline","score":0,"run_count":9}
+{"timestamp":"2025-12-25T16:40:39.352Z","eval_name":"Example: Basic scorer test","score":1,"run_count":54}
+{"timestamp":"2025-12-25T16:40:39.460Z","eval_name":"Example: Basic scorer test","score":1,"run_count":55}
+{"timestamp":"2025-12-25T16:40:39.572Z","eval_name":"Example: Basic scorer test","score":1,"run_count":56}
+{"timestamp":"2025-12-25T16:40:39.816Z","eval_name":"Example: Basic scorer test","score":1,"run_count":57}
+{"timestamp":"2025-12-25T16:40:39.947Z","eval_name":"Example: Basic scorer test","score":1,"run_count":58}
+{"timestamp":"2025-12-25T16:40:40.084Z","eval_name":"Example: Basic scorer test","score":1,"run_count":59}
+{"timestamp":"2025-12-25T16:40:40.202Z","eval_name":"Example: Basic scorer test","score":1,"run_count":60}
+{"timestamp":"2025-12-25T16:43:12.851Z","eval_name":"Example: Basic scorer test","score":1,"run_count":61}
+{"timestamp":"2025-12-25T16:43:43.041Z","eval_name":"Coordinator Discipline - Synthetic Fixtures","score":0.6354444444444443,"run_count":10}
+{"timestamp":"2025-12-25T16:43:43.041Z","eval_name":"Coordinator Discipline - Real Sessions","score":0,"run_count":10}
+{"timestamp":"2025-12-25T16:43:43.041Z","eval_name":"Coordinator Discipline - Perfect vs Bad","score":0.5416666666666666,"run_count":10}
+{"timestamp":"2025-12-25T16:43:43.042Z","eval_name":"Swarm Decomposition Quality","score":0.6909722222222222,"run_count":10}
+{"timestamp":"2025-12-25T16:43:43.042Z","eval_name":"Decomposition Edge Cases","score":0.775,"run_count":10}
+{"timestamp":"2025-12-25T16:43:43.042Z","eval_name":"Compaction Hook Coordinator Resumption","score":0.95,"run_count":10}
+{"timestamp":"2025-12-25T16:43:43.042Z","eval_name":"Epic ID Specificity","score":0.5,"run_count":10}
+{"timestamp":"2025-12-25T16:43:43.042Z","eval_name":"No False Positives","score":1,"run_count":10}
+{"timestamp":"2025-12-25T16:43:43.042Z","eval_name":"Example: Basic scorer test","score":1,"run_count":62}
+{"timestamp":"2025-12-25T16:43:43.042Z","eval_name":"Perfect Prompt Scores 100%","score":1,"run_count":10}
+{"timestamp":"2025-12-25T16:43:43.042Z","eval_name":"Placeholder Detection","score":0,"run_count":10}
+{"timestamp":"2025-12-25T16:43:43.042Z","eval_name":"Generic Instructions Fail","score":0,"run_count":10}
+{"timestamp":"2025-12-25T16:43:43.042Z","eval_name":"First Tool Discipline","score":0,"run_count":10}
+{"timestamp":"2025-12-25T16:43:43.042Z","eval_name":"Coordinator Resists Direct Implementation","score":1,"run_count":10}
+{"timestamp":"2025-12-25T16:43:43.042Z","eval_name":"Compaction Prompt Quality","score":0.6342857142857142,"run_count":10}
+{"timestamp":"2025-12-25T16:43:43.043Z","eval_name":"Coordinator Behavior After Compaction","score":1,"run_count":10}
+{"timestamp":"2025-12-25T16:44:12.471Z","eval_name":"Coordinator Discipline - Synthetic Fixtures","score":0.6354444444444443,"run_count":11}
+{"timestamp":"2025-12-25T16:44:12.471Z","eval_name":"Coordinator Discipline - Real Sessions","score":0,"run_count":11}
+{"timestamp":"2025-12-25T16:44:12.471Z","eval_name":"Coordinator Discipline - Perfect vs Bad","score":0.5416666666666666,"run_count":11}
+{"timestamp":"2025-12-25T16:44:12.471Z","eval_name":"Swarm Decomposition Quality","score":0.6720833333333333,"run_count":11}
+{"timestamp":"2025-12-25T16:44:12.471Z","eval_name":"Decomposition Edge Cases","score":0.775,"run_count":11}
+{"timestamp":"2025-12-25T16:44:12.471Z","eval_name":"Example: Basic scorer test","score":1,"run_count":63}
+{"timestamp":"2025-12-25T16:44:12.471Z","eval_name":"Compaction Hook Coordinator Resumption","score":0.95,"run_count":11}
+{"timestamp":"2025-12-25T16:44:12.472Z","eval_name":"Epic ID Specificity","score":0.5,"run_count":11}
+{"timestamp":"2025-12-25T16:44:12.472Z","eval_name":"No False Positives","score":1,"run_count":11}
+{"timestamp":"2025-12-25T16:44:12.472Z","eval_name":"Coordinator Behavior After Compaction","score":0.9796875,"run_count":11}
+{"timestamp":"2025-12-25T16:44:12.472Z","eval_name":"Coordinator Resists Direct Implementation","score":1,"run_count":11}
+{"timestamp":"2025-12-25T16:44:12.472Z","eval_name":"Compaction Prompt Quality","score":0.6342857142857142,"run_count":11}
+{"timestamp":"2025-12-25T16:44:12.472Z","eval_name":"Perfect Prompt Scores 100%","score":1,"run_count":11}
+{"timestamp":"2025-12-25T16:44:12.472Z","eval_name":"Placeholder Detection","score":0,"run_count":11}
+{"timestamp":"2025-12-25T16:44:12.472Z","eval_name":"Generic Instructions Fail","score":0,"run_count":11}
+{"timestamp":"2025-12-25T16:44:12.472Z","eval_name":"First Tool Discipline","score":0,"run_count":11}
+{"timestamp":"2025-12-25T16:49:55.548Z","eval_name":"Coordinator Discipline - Synthetic Fixtures","score":0.6354444444444443,"run_count":12}
+{"timestamp":"2025-12-25T16:49:55.549Z","eval_name":"Coordinator Discipline - Real Sessions","score":0,"run_count":12}
+{"timestamp":"2025-12-25T16:49:55.555Z","eval_name":"Coordinator Discipline - Perfect vs Bad","score":0.5416666666666666,"run_count":12}
+{"timestamp":"2025-12-25T16:49:55.555Z","eval_name":"Decomposition Edge Cases","score":0.775,"run_count":12}
+{"timestamp":"2025-12-25T16:49:55.555Z","eval_name":"Swarm Decomposition Quality","score":0.7001388888888888,"run_count":12}
+{"timestamp":"2025-12-25T16:49:55.555Z","eval_name":"Compaction Hook Coordinator Resumption","score":0.95,"run_count":12}
+{"timestamp":"2025-12-25T16:49:55.556Z","eval_name":"Epic ID Specificity","score":0.5,"run_count":12}
+{"timestamp":"2025-12-25T16:49:55.557Z","eval_name":"No False Positives","score":1,"run_count":12}
+{"timestamp":"2025-12-25T16:49:55.557Z","eval_name":"Example: Basic scorer test","score":1,"run_count":64}
+{"timestamp":"2025-12-25T16:49:55.557Z","eval_name":"Generic Instructions Fail","score":0,"run_count":12}
+{"timestamp":"2025-12-25T16:49:55.557Z","eval_name":"First Tool Discipline","score":0,"run_count":12}
+{"timestamp":"2025-12-25T16:49:55.561Z","eval_name":"Compaction Prompt Quality","score":0.6342857142857142,"run_count":12}
+{"timestamp":"2025-12-25T16:49:55.561Z","eval_name":"Perfect Prompt Scores 100%","score":1,"run_count":12}
+{"timestamp":"2025-12-25T16:49:55.561Z","eval_name":"Placeholder Detection","score":0,"run_count":12}
+{"timestamp":"2025-12-25T16:49:55.561Z","eval_name":"Coordinator Behavior After Compaction","score":1,"run_count":12}
+{"timestamp":"2025-12-25T16:49:55.565Z","eval_name":"Coordinator Resists Direct Implementation","score":0.9375,"run_count":12}
+{"timestamp":"2025-12-25T16:49:55.697Z","eval_name":"Example: Basic scorer test","score":1,"run_count":65}
+{"timestamp":"2025-12-25T16:49:55.813Z","eval_name":"Example: Basic scorer test","score":1,"run_count":66}
+{"timestamp":"2025-12-25T16:49:55.934Z","eval_name":"Example: Basic scorer test","score":1,"run_count":67}
+{"timestamp":"2025-12-25T16:49:56.178Z","eval_name":"Example: Basic scorer test","score":1,"run_count":68}
+{"timestamp":"2025-12-25T16:49:56.327Z","eval_name":"Example: Basic scorer test","score":1,"run_count":69}
+{"timestamp":"2025-12-25T16:49:56.446Z","eval_name":"Example: Basic scorer test","score":1,"run_count":70}
+{"timestamp":"2025-12-25T16:49:56.556Z","eval_name":"Example: Basic scorer test","score":1,"run_count":71}
+{"timestamp":"2025-12-25T17:06:10.610Z","eval_name":"Coordinator Discipline - Synthetic Fixtures","score":0.6354444444444443,"run_count":13}
+{"timestamp":"2025-12-25T17:06:10.610Z","eval_name":"Coordinator Discipline - Real Sessions","score":0,"run_count":13}
+{"timestamp":"2025-12-25T17:06:10.610Z","eval_name":"Coordinator Discipline - Perfect vs Bad","score":0.5416666666666666,"run_count":13}
+{"timestamp":"2025-12-25T17:06:10.610Z","eval_name":"Decomposition Edge Cases","score":0.775,"run_count":13}
+{"timestamp":"2025-12-25T17:06:10.610Z","eval_name":"Swarm Decomposition Quality","score":0.6847222222222221,"run_count":13}
+{"timestamp":"2025-12-25T17:06:10.610Z","eval_name":"Compaction Hook Coordinator Resumption","score":0.95,"run_count":13}
+{"timestamp":"2025-12-25T17:06:10.610Z","eval_name":"Epic ID Specificity","score":0.5,"run_count":13}
+{"timestamp":"2025-12-25T17:06:10.610Z","eval_name":"No False Positives","score":1,"run_count":13}
+{"timestamp":"2025-12-25T17:06:10.611Z","eval_name":"Example: Basic scorer test","score":1,"run_count":72}
+{"timestamp":"2025-12-25T17:06:10.611Z","eval_name":"Perfect Prompt Scores 100%","score":1,"run_count":13}
+{"timestamp":"2025-12-25T17:06:10.611Z","eval_name":"Placeholder Detection","score":0,"run_count":13}
+{"timestamp":"2025-12-25T17:06:10.611Z","eval_name":"Generic Instructions Fail","score":0,"run_count":13}
+{"timestamp":"2025-12-25T17:06:10.611Z","eval_name":"First Tool Discipline","score":0,"run_count":13}
+{"timestamp":"2025-12-25T17:06:10.611Z","eval_name":"Compaction Prompt Quality","score":0.6342857142857142,"run_count":13}
+{"timestamp":"2025-12-25T17:06:10.611Z","eval_name":"Coordinator Behavior After Compaction","score":0.8645833333333333,"run_count":13}
+{"timestamp":"2025-12-25T17:06:10.611Z","eval_name":"Coordinator Resists Direct Implementation","score":0.9375,"run_count":13}
+{"timestamp":"2025-12-25T18:58:44.923Z","eval_name":"Coordinator Discipline - Synthetic Fixtures","score":0.6354444444444443,"run_count":14}
+{"timestamp":"2025-12-25T18:58:44.923Z","eval_name":"Coordinator Discipline - Real Sessions","score":0,"run_count":14}
+{"timestamp":"2025-12-25T18:58:44.923Z","eval_name":"Coordinator Discipline - Perfect vs Bad","score":0.5416666666666666,"run_count":14}
+{"timestamp":"2025-12-25T18:58:44.924Z","eval_name":"Swarm Decomposition Quality","score":0.7095833333333333,"run_count":14}
+{"timestamp":"2025-12-25T18:58:44.924Z","eval_name":"Decomposition Edge Cases","score":0.775,"run_count":14}
+{"timestamp":"2025-12-25T18:58:44.924Z","eval_name":"Compaction Hook Coordinator Resumption","score":0.95,"run_count":14}
+{"timestamp":"2025-12-25T18:58:44.924Z","eval_name":"Epic ID Specificity","score":0.5,"run_count":14}
+{"timestamp":"2025-12-25T18:58:44.924Z","eval_name":"No False Positives","score":1,"run_count":14}
+{"timestamp":"2025-12-25T18:58:44.924Z","eval_name":"Example: Basic scorer test","score":1,"run_count":73}
+{"timestamp":"2025-12-25T18:58:44.924Z","eval_name":"Generic Instructions Fail","score":0,"run_count":14}
+{"timestamp":"2025-12-25T18:58:44.924Z","eval_name":"First Tool Discipline","score":0,"run_count":14}
+{"timestamp":"2025-12-25T18:58:44.924Z","eval_name":"Compaction Prompt Quality","score":0.6342857142857142,"run_count":14}
+{"timestamp":"2025-12-25T18:58:44.925Z","eval_name":"Perfect Prompt Scores 100%","score":1,"run_count":14}
+{"timestamp":"2025-12-25T18:58:44.925Z","eval_name":"Placeholder Detection","score":0,"run_count":14}
+{"timestamp":"2025-12-25T18:58:44.925Z","eval_name":"Coordinator Behavior After Compaction","score":0.9375,"run_count":14}
+{"timestamp":"2025-12-25T18:58:44.925Z","eval_name":"Coordinator Resists Direct Implementation","score":1,"run_count":14}
+{"timestamp":"2025-12-25T18:59:58.928Z","eval_name":"Coordinator Discipline - Synthetic Fixtures","score":0.6354444444444443,"run_count":15}
+{"timestamp":"2025-12-25T18:59:58.929Z","eval_name":"Coordinator Discipline - Real Sessions","score":0,"run_count":15}
+{"timestamp":"2025-12-25T18:59:58.929Z","eval_name":"Coordinator Discipline - Perfect vs Bad","score":0.5416666666666666,"run_count":15}
+{"timestamp":"2025-12-25T18:59:58.929Z","eval_name":"Swarm Decomposition Quality","score":0.6944444444444443,"run_count":15}
+{"timestamp":"2025-12-25T18:59:58.929Z","eval_name":"Decomposition Edge Cases","score":0.775,"run_count":15}
+{"timestamp":"2025-12-25T18:59:58.929Z","eval_name":"Compaction Hook Coordinator Resumption","score":0.95,"run_count":15}
+{"timestamp":"2025-12-25T18:59:58.929Z","eval_name":"Epic ID Specificity","score":0.5,"run_count":15}
+{"timestamp":"2025-12-25T18:59:58.929Z","eval_name":"No False Positives","score":1,"run_count":15}
+{"timestamp":"2025-12-25T18:59:58.929Z","eval_name":"Example: Basic scorer test","score":1,"run_count":74}
+{"timestamp":"2025-12-25T18:59:58.929Z","eval_name":"Compaction Prompt Quality","score":0.6342857142857142,"run_count":15}
+{"timestamp":"2025-12-25T18:59:58.929Z","eval_name":"Perfect Prompt Scores 100%","score":1,"run_count":15}
+{"timestamp":"2025-12-25T18:59:58.930Z","eval_name":"Placeholder Detection","score":0,"run_count":15}
+{"timestamp":"2025-12-25T18:59:58.930Z","eval_name":"Generic Instructions Fail","score":0,"run_count":15}
+{"timestamp":"2025-12-25T18:59:58.930Z","eval_name":"First Tool Discipline","score":0,"run_count":15}
+{"timestamp":"2025-12-25T18:59:58.930Z","eval_name":"Coordinator Behavior After Compaction","score":0.9171875,"run_count":15}
+{"timestamp":"2025-12-25T18:59:58.930Z","eval_name":"Coordinator Resists Direct Implementation","score":1,"run_count":15}
+{"timestamp":"2025-12-25T19:00:48.709Z","eval_name":"Coordinator Discipline - Synthetic Fixtures","score":0.6354444444444443,"run_count":16}
+{"timestamp":"2025-12-25T19:00:48.709Z","eval_name":"Coordinator Discipline - Real Sessions","score":0,"run_count":16}
+{"timestamp":"2025-12-25T19:00:48.710Z","eval_name":"Coordinator Discipline - Perfect vs Bad","score":0.5416666666666666,"run_count":16}
+{"timestamp":"2025-12-25T19:00:48.710Z","eval_name":"Swarm Decomposition Quality","score":0.5464583333333334,"run_count":16}
+{"timestamp":"2025-12-25T19:00:48.710Z","eval_name":"Decomposition Edge Cases","score":0.775,"run_count":16}
+{"timestamp":"2025-12-25T19:00:48.710Z","eval_name":"Compaction Hook Coordinator Resumption","score":0.95,"run_count":16}
+{"timestamp":"2025-12-25T19:00:48.710Z","eval_name":"Epic ID Specificity","score":0.5,"run_count":16}
+{"timestamp":"2025-12-25T19:00:48.710Z","eval_name":"No False Positives","score":1,"run_count":16}
+{"timestamp":"2025-12-25T19:00:48.710Z","eval_name":"Example: Basic scorer test","score":1,"run_count":75}
+{"timestamp":"2025-12-25T19:00:48.710Z","eval_name":"Perfect Prompt Scores 100%","score":1,"run_count":16}
+{"timestamp":"2025-12-25T19:00:48.711Z","eval_name":"Placeholder Detection","score":0,"run_count":16}
+{"timestamp":"2025-12-25T19:00:48.711Z","eval_name":"Generic Instructions Fail","score":0,"run_count":16}
+{"timestamp":"2025-12-25T19:00:48.711Z","eval_name":"First Tool Discipline","score":0,"run_count":16}
+{"timestamp":"2025-12-25T19:00:48.711Z","eval_name":"Compaction Prompt Quality","score":0.6342857142857142,"run_count":16}
+{"timestamp":"2025-12-25T19:00:48.711Z","eval_name":"Coordinator Behavior After Compaction","score":1,"run_count":16}
+{"timestamp":"2025-12-25T19:00:48.711Z","eval_name":"Coordinator Resists Direct Implementation","score":1,"run_count":16}

package/.turbo/turbo-build.log

@@ -1,9 +1,9 @@
-$ bun build ./src/index.ts --outdir ./dist --target node --external @electric-sql/pglite --external swarm-mail && bun build ./src/plugin.ts --outfile ./dist/plugin.js --target node --external @electric-sql/pglite --external swarm-mail && tsc
-Bundled 917 modules in 237ms
+$ bun build ./src/index.ts --outdir ./dist --target node --external @electric-sql/pglite --external swarm-mail --external vitest --external @vitest/ui --external lightningcss && bun build ./src/plugin.ts --outfile ./dist/plugin.js --target node --external @electric-sql/pglite --external swarm-mail --external vitest --external @vitest/ui --external lightningcss && tsc
+Bundled 1348 modules in 205ms
 
-  index.js  2.16 MB  (entry point)
+  index.js  4.33 MB  (entry point)
 
-Bundled 918 modules in 82ms
+Bundled 1349 modules in 196ms
 
-  plugin.js  2.12 MB  (entry point)
+  plugin.js  4.30 MB  (entry point)
 

package/CHANGELOG.md

@@ -1,5 +1,212 @@
 # opencode-swarm-plugin
 
+## 0.42.0
+
+### Minor Changes
+
+- [`a79e04b`](https://github.com/joelhooks/swarm-tools/commit/a79e04b1bb3b40c09c5265b5d11739864799e4e2) Thanks [@joelhooks](https://github.com/joelhooks)! - ## 🔭 Swarm Observability: See What Your Bees Are Doing
+
+  > "Observability is about instrumenting your system in a way that ensures sufficient information about a system's runtime is collected and analyzed so that when something goes wrong, it can help you understand why."
+  > — Chip Huyen, _AI Engineering_
+
+  New CLI commands to understand swarm health and history:
+
+  ### `swarm stats`
+
+  ```
+  ┌─────────────────────────────────────────┐
+  │        🐝  SWARM STATISTICS  🐝         │
+  ├─────────────────────────────────────────┤
+  │ Total Swarms: 42   Success: 87%         │
+  │ Avg Duration: 4.2min                    │
+  ├─────────────────────────────────────────┤
+  │ BY STRATEGY                             │
+  │ ├─ file-based      92% (23/25)          │
+  │ ├─ feature-based   78% (14/18)          │
+  │ ├─ risk-based      67% (2/3)            │
+  ├─────────────────────────────────────────┤
+  │ COORDINATOR HEALTH                      │
+  │ Violation Rate:   2%                    │
+  │ Spawn Efficiency: 94%                   │
+  │ Review Rate:      88%                   │
+  └─────────────────────────────────────────┘
+  ```
+
+  Options: `--since 24h/7d/30d`, `--json`
+
+  ### `swarm history`
+
+  Timeline of recent swarm activity with filtering:
+
+  - `--status success/failed/in_progress`
+  - `--strategy file-based/feature-based/risk-based`
+  - `--verbose` for subtask details
+
+  ### Prompt Insights Integration
+
+  Coordinators and workers now receive injected insights from past swarm outcomes:
+
+  - Strategy success rates as markdown tables
+  - Anti-pattern warnings for low-success strategies
+  - File/domain-specific learnings from semantic memory
+
+  This creates a feedback loop where swarms learn from their own history.
+
+  ### Also in this release
+
+  - **swarm-dashboard** (WIP): React/Vite visualizer scaffold
+  - **ADR-006**: Swarm PTY decision document
+  - **CI fix**: Smarter changeset detection prevents empty PR errors
+
+### Patch Changes
+
+- Updated dependencies [[`a79e04b`](https://github.com/joelhooks/swarm-tools/commit/a79e04b1bb3b40c09c5265b5d11739864799e4e2)]:
+  - swarm-mail@1.5.4
+
+## 0.41.0
+
+### Minor Changes
+
+- [`179b3f0`](https://github.com/joelhooks/swarm-tools/commit/179b3f0e49c7959f8d754c1274d301d0b3845a79) Thanks [@joelhooks](https://github.com/joelhooks)! - ## 🐝 Compaction Prompt Now Speaks Swarm
+
+  > _"Memory is essential for communication: we recall past interactions, infer preferences, and construct evolving mental models of those we engage with."_
+  > — Mem0: Building Production-Ready AI Agents with Scalable Long-Term Memory
+
+  When context compacts mid-swarm, coordinators were waking up confused. They had state information but no protocol guidance. Now the compaction prompt includes a condensed version of the swarm command template.
+
+  **What's New:**
+
+  The `SWARM_COMPACTION_CONTEXT` now includes:
+
+  1. **What Good Looks Like** - Behavioral examples showing ideal coordinator behavior
+
+     - ✅ Spawned researcher for unfamiliar tech → got summary → stored in semantic-memory
+     - ✅ Checked inbox every 5-10 minutes → caught blocked worker → unblocked in 2min
+     - ❌ Called context7 directly → dumped 50KB → context exhaustion
+
+  2. **Mandatory Behaviors Checklist** - Post-compaction protocol
+     - Inbox monitoring (every 5-10 min with intervention triggers)
+     - Skill loading (before spawning workers)
+     - Worker review (after every worker returns, 3-strike rule)
+     - Research spawning (never call context7/pdf-brain directly)
+
+  **Why This Matters:**
+
+  Coordinators resuming from compaction now have:
+
+  - Clear behavioral guidance (not just state)
+  - Actionable tool call examples
+  - Anti-patterns to avoid
+  - The same protocol as fresh `/swarm` invocations
+
+  **Backward Compatible:** Existing compaction hooks continue to work. This adds guidance, doesn't change the hook signature.
+
+### Patch Changes
+
+- [`3e7c126`](https://github.com/joelhooks/swarm-tools/commit/3e7c126b11aa6ad909ebcb2ab3cf77883f9acfe4) Thanks [@joelhooks](https://github.com/joelhooks)! - ## 🧪 Bulletproof Test Suite
+
+  > "Setting up our tests to run synchronously and using mocking libraries will greatly speed up our testing"
+  > — ng-book
+
+  Fixed test isolation issues that caused 19 tests to fail when run together but pass in isolation.
+
+  ### The Culprits
+
+  **1. Global fetch pollution** (`ollama.test.ts`)
+
+  ```typescript
+  // BEFORE: Replaced global.fetch, never restored it
+  global.fetch = mockFetch;
+
+  // AFTER: Save and restore
+  const originalFetch = global.fetch;
+  afterEach(() => {
+    global.fetch = originalFetch;
+  });
+  ```
+
+  **2. Port conflicts** (`durable-server.test.ts`)
+
+  - Tests used hardcoded ports (4483, 4484, 4485)
+  - Parallel test runs fought over the same ports
+  - Fixed: Use `port: 0` for OS-assigned ports, made `server.url` a getter
+
+  **3. AI SDK schema incompatibility** (`memory-operations.ts`)
+
+  - `z.discriminatedUnion` creates `oneOf` at top level
+  - Anthropic API requires `type: object` at top level
+  - Fixed: Flat object schema with optional fields
+
+  ### Test Stats
+
+  ```
+  Before: 19 failures when run together
+  After:  0 failures, 1406 tests pass
+  ```
+
+  ### Files Changed
+
+  - `src/memory/ollama.test.ts` - Restore global.fetch after each test
+  - `src/streams/durable-server.ts` - Dynamic port getter
+  - `src/streams/durable-server.test.ts` - Use port 0, rewrite for isolation
+  - `src/memory/memory-operations.ts` - Flat schema for Anthropic compatibility
+  - Renamed `memory-operations.test.ts` → `memory-operations.integration.test.ts`
+
+- Updated dependencies [[`3e7c126`](https://github.com/joelhooks/swarm-tools/commit/3e7c126b11aa6ad909ebcb2ab3cf77883f9acfe4)]:
+  - swarm-mail@1.5.3
+
+## 0.40.0
+
+### Minor Changes
+
+- [`948e031`](https://github.com/joelhooks/swarm-tools/commit/948e0318fe5e2c1a5d695a56533fc2a2a7753887) Thanks [@joelhooks](https://github.com/joelhooks)! - ## 🔭 Observability Swarm: See What the Bees Are Doing
+
+  > "The unexamined swarm is not worth coordinating." — Socrates, probably
+
+  Four parallel workers descended on the observability stack and emerged victorious. The compaction hook no longer runs in darkness, coordinator sessions are now viewable, and the docs finally explain what all those JSONL files are for.
+
+  ### What's New
+
+  **Compaction Observability** (`src/compaction-observability.ts`)
+
+  - Metrics collector tracks phases: START → GATHER → DETECT → INJECT → COMPLETE
+  - Pattern extraction/skipping with reasons ("why didn't this get captured?")
+  - Timing breakdown per phase (analysis vs extraction vs storage)
+  - 15 tests (11 unit + 4 integration)
+
+  **`swarm log sessions` CLI**
+
+  - `swarm log sessions` — list all captured coordinator sessions
+  - `swarm log sessions <id>` — view events for a session (partial ID matching)
+  - `swarm log sessions --latest` — quick access to most recent
+  - `--type`, `--since`, `--limit`, `--json` filters
+  - 64 tests covering parsing, listing, filtering
+
+  **Coordinator Observability Docs**
+
+  - AGENTS.md: overview with quick commands
+  - evals/README.md: deep dive with ASCII flow diagrams, event type reference, JSONL examples, jq recipes
+
+  **Research: Coordinator Prompt Eval** (`.hive/analysis/coordinator-prompt-eval-research.md`)
+
+  - 26KB analysis of prompt iteration strategies
+  - Recommends: versioning + evalite (defer LLM-as-Judge to v0.34+)
+  - Implementation plan with effort estimates
+
+  ### The Observability Story
+
+  ```
+  CAPTURE ──────────► VIEW ──────────► SCORE
+  (eval-capture.ts)   (swarm log       (coordinator
+                       sessions)        evals)
+  ```
+
+  Now you can answer:
+
+  - "What did the last 10 compaction runs extract?"
+  - "Why didn't this pattern get captured?"
+  - "Which coordinator sessions had violations?"
+
 ## 0.39.1
 
 ### Patch Changes

package/README.md

@@ -5,6 +5,8 @@
 **🌐 Website:** [swarmtools.ai](https://swarmtools.ai)  
 **📚 Full Documentation:** [swarmtools.ai/docs](https://swarmtools.ai/docs)
 
+[![Eval Gate](https://github.com/joelhooks/opencode-swarm-plugin/actions/workflows/eval-gate.yml/badge.svg)](https://github.com/joelhooks/opencode-swarm-plugin/actions/workflows/eval-gate.yml)
+
 ```
  ███████╗██╗    ██╗ █████╗ ██████╗ ███╗   ███╗
  ██╔════╝██║    ██║██╔══██╗██╔══██╗████╗ ████║