{ "title": "Self-Improving Skill Loop | Actual artifact output", "recordedAt": "2026-09-28T19:22:00Z", "records": [ { "command": [ "python3.12", "demo.py" ], "cwd": "/tmp/aiefs-recorded-artifacts/self-improving-skill-loop/solution", "exitCode": 0, "output": "{\n \"candidate\": {\n \"schema_version\": 1,\n \"dataset_sha256\": \"c4fde1e162290765a79b5f3a45f8f95286026c95b24f34981b815b6bea6d5558\",\n \"development_cases\": [\n {\n \"id\": \"dev-1\",\n \"group\": \"invoice-june\",\n \"text\": \"invoice charged twice\",\n \"label\": \"billing\",\n \"source\": \"local:invoice-june\"\n },\n {\n \"id\": \"dev-2\",\n \"group\": \"invoice-july\",\n \"text\": \"invoice refund pending\",\n \"label\": \"billing\",\n \"source\": \"local:invoice-july\"\n },\n {\n \"id\": \"dev-3\",\n \"group\": \"login-may\",\n \"text\": \"password reset link expired\",\n \"label\": \"access\",\n \"source\": \"local:login-may\"\n },\n {\n \"id\": \"dev-4\",\n \"group\": \"login-june\",\n \"text\": \"password reset email missing\",\n \"label\": \"access\",\n \"source\": \"local:login-june\"\n }\n ],\n \"holdout_cases\": [\n {\n \"id\": \"hold-1\",\n \"group\": \"invoice-august\",\n \"text\": \"invoice tax needs correction\",\n \"label\": \"billing\",\n \"source\": \"local:invoice-august\"\n },\n {\n \"id\": \"hold-2\",\n \"group\": \"login-july\",\n \"text\": \"password rejected after rotation\",\n \"label\": \"access\",\n \"source\": \"local:login-july\"\n }\n ],\n \"audit\": {\n \"content_leaks\": [],\n \"group_leaks\": [],\n \"usable\": true\n },\n \"development_ids\": [\n \"dev-1\",\n \"dev-2\",\n \"dev-3\",\n \"dev-4\"\n ],\n \"holdout_ids\": [\n \"hold-1\",\n \"hold-2\"\n ],\n \"rules\": [\n {\n \"terms\": [\n \"invoice\"\n ],\n \"label\": \"billing\"\n },\n {\n \"terms\": [\n \"password\"\n ],\n \"label\": \"access\"\n },\n {\n \"terms\": [\n \"reset\"\n ],\n \"label\": \"access\"\n }\n ],\n \"development\": {\n \"total\": 4,\n \"accuracy\": 1.0,\n \"errors\": [],\n \"rows\": [\n {\n \"id\": \"dev-1\",\n \"expected\": \"billing\",\n \"predicted\": \"billing\"\n },\n {\n \"id\": \"dev-2\",\n \"expected\": \"billing\",\n \"predicted\": \"billing\"\n },\n {\n \"id\": \"dev-3\",\n \"expected\": \"access\",\n \"predicted\": \"access\"\n },\n {\n \"id\": \"dev-4\",\n \"expected\": \"access\",\n \"predicted\": \"access\"\n }\n ]\n },\n \"holdout\": {\n \"total\": 2,\n \"accuracy\": 1.0,\n \"errors\": [],\n \"rows\": [\n {\n \"id\": \"hold-1\",\n \"expected\": \"billing\",\n \"predicted\": \"billing\"\n },\n {\n \"id\": \"hold-2\",\n \"expected\": \"access\",\n \"predicted\": \"access\"\n }\n ]\n },\n \"gate\": {\n \"promote\": true,\n \"gain\": 1.0,\n \"regressions\": [],\n \"candidate_sha256\": \"62cf483e8839424675369c11e5ab0057836b031e81aaa002edac202660e45f27\",\n \"baseline\": 0.0,\n \"candidate\": 1.0\n }\n },\n \"promotion\": \"not requested; inspect candidate digest before approval\"\n}\n", "durationMs": 43 } ] }