{ "created_at": "2026-11-12T07:59:12Z", "result": { "engine": "rnd_change", "limits": "Reviews supplied numbers and test results. It does not re-run the experiment, rebuild the code, or prove the metric measures what the team thinks. The run is Ed25519-signed; pin the issuer key out of band.", "module": "cain_a75_engines/rndchange/engine.py", "profile": "parser_roundtrip_fix", "run": { "label": "Recomputes an autonomous R&D change claim: diff and authority binding, preregistration, CI-clearing gain, a test that failed then passes, regressions, held-out evidence and claim honesty. It does not re-run the experiment or the code.", "public_key_b64": "UuwlUy36flMvwAXAkmsNwprs1LullQjv/4CVuce90ag=", "records": [ { "checks": [ { "detail": "the diff hash covers exactly these files", "id": "binding.diff_hash", "ok": true }, { "detail": "linked 'exp-2026-11-09-parser'", "id": "binding.experiment_linked", "ok": true }, { "detail": "the built artifact is named by hash", "id": "binding.artifact_hash", "ok": true }, { "detail": "named principal", "id": "authority.principal", "ok": true }, { "detail": "scope ['change', 'experiment']", "id": "authority.scope", "ok": true }, { "detail": "expires 2000.0 > now 1000.0", "id": "authority.not_expired", "ok": true }, { "detail": "the hypothesis and metric were fixed before the run", "id": "experiment.preregistered", "ok": true }, { "detail": "30 >= 30", "id": "experiment.trials", "ok": true }, { "detail": "seed 7", "id": "experiment.seed", "ok": true }, { "detail": "direction 'higher_better'", "id": "experiment.direction", "ok": true }, { "detail": "gain 0.09 > 95% CI 0.02", "id": "experiment.clears_ci", "ok": true }, { "detail": "metric 'heldout_accuracy'", "id": "experiment.metric_matches", "ok": true }, { "detail": "failed then passed: ['test_parser_roundtrip']", "id": "tests.one_fix_demonstrated", "ok": true }, { "detail": "passed then failed: []", "id": "tests.no_regression", "ok": true }, { "detail": "result 'pass'", "id": "held_out.present", "ok": true }, { "detail": "held-out was not used to pick the change", "id": "held_out.not_used_for_selection", "ok": true }, { "detail": "the held-out check passes", "id": "held_out.passed", "ok": true }, { "detail": "claimed True, computed True", "id": "claims.held_out_passed", "ok": true }, { "detail": "claimed True, computed True", "id": "claims.improves_metric", "ok": true }, { "detail": "claimed True, computed True", "id": "claims.no_regression", "ok": true }, { "detail": "claimed True, computed True", "id": "claims.tests_pass", "ok": true } ], "created_at": "2026-11-12T07:59:12Z", "engine": "a75-009/1.0.0", "extra": { "ci95": 0.02, "computed": { "held_out_passed": true, "improves_metric": true, "no_regression": true, "tests_pass": true }, "fixed_tests": [ "test_parser_roundtrip" ], "gain": 0.09, "regressed_tests": [] }, "failed": [], "label": "Recomputes an autonomous R&D change claim: diff and authority binding, preregistration, CI-clearing gain, a test that failed then passes, regressions, held-out evidence and claim honesty. It does not re-run the experiment or the code.", "prev_record_hash": "9c13847d31287273122ecfae44179078476d6bb67d2961038e5d408afefa3c31", "public_key_b64": "UuwlUy36flMvwAXAkmsNwprs1LullQjv/4CVuce90ag=", "record_hash": "d96760b1c201472ef65451940eb4026f56b47c198985c74cdda7551e6e545e6c", "schema": "cain42.a75.rnd.record.v1", "signature_b64": "kDlsm2fbXyHVrcGZ4rNaz+laQBuk6gxJaOWIAKPhMPwmqM1sESF0u7ep3DA9E/0YXyANPoMWD7hpkVhwUHmKCw==", "sku": "CAIN-A75-009", "subject": { "authority": { "expires_at": 2000.0, "now": 1000.0, "principal": "agent:rnd-loop", "scope": [ "change", "experiment" ] }, "change": { "artifact_sha256": "aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa", "diff_sha256": "8d8bf13a3b6725e517b757487ae59448385108d3b6245fb0e2a0873677733569", "experiment_id": "exp-2026-11-09-parser", "files": [ { "after_sha256": "cccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccc", "before_sha256": "bbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbb", "path": "src/parser.py" } ], "id": "change-parser-roundtrip", "rationale": "Fix the parser dropping the final token on round-trip." }, "claims": { "held_out_passed": true, "improves_metric": true, "no_regression": true, "tests_pass": true }, "experiment": { "baseline_value": 0.72, "candidate_value": 0.81, "ci95": 0.02, "direction": "higher_better", "hypothesis": "Round-trip preserves every token.", "id": "exp-2026-11-09-parser", "metric": "heldout_accuracy", "min_trials": 30, "preregistered": true, "seed": 7, "trials": 30 }, "held_out": { "name": "heldout-corpus-v3", "result": "pass", "used_for_selection": false }, "schema": "cain42.a75.rnd.change.v1", "tests": [ { "after": "pass", "before": "fail", "name": "test_parser_roundtrip" }, { "after": "pass", "before": "pass", "name": "test_cli_help" } ] }, "subject_sha256": "307fd0b1baf02784b42c35fd1096aba35114a55b4260b441da31c1bd2ac94b82", "verdict": "ACCEPT" } ], "run_hash": "e7edef2f94c43b7ff63b9e78a370f51987ee476f7412be24eaa61ee9bc275e90", "schema": "cain42.a75.rnd.run.v1", "sku": "CAIN-A75-009", "verdict": "VERIFIED", "verdicts": [ "ACCEPT" ] }, "scenarios": [ { "ci95": 0.02, "fixed_tests": [ "test_parser_roundtrip" ], "gain": 0.09, "name": "parser_roundtrip_change", "run_hash": "e7edef2f94c43b7ff63b9e78a370f51987ee476f7412be24eaa61ee9bc275e90", "verdict": "VERIFIED" }, { "expected": "ACCEPT", "name": "honest_improvement", "verdict": "ACCEPT" }, { "expected": "REJECT", "name": "gain_within_confidence_interval", "verdict": "REJECT" }, { "expected": "REJECT", "name": "no_failing_test_that_now_passes", "verdict": "REJECT" }, { "expected": "REJECT", "name": "regression_introduced", "verdict": "REJECT" }, { "expected": "REJECT", "name": "not_preregistered", "verdict": "REJECT" }, { "expected": "REJECT", "name": "too_few_trials", "verdict": "REJECT" }, { "expected": "REJECT", "name": "seed_missing", "verdict": "REJECT" }, { "expected": "REJECT", "name": "authorization_expired", "verdict": "REJECT" }, { "expected": "REJECT", "name": "scope_does_not_cover_change", "verdict": "REJECT" }, { "expected": "REJECT", "name": "diff_hash_mismatch", "verdict": "REJECT" }, { "expected": "REJECT", "name": "rationale_missing", "verdict": "REJECT" }, { "expected": "REJECT", "name": "held_out_missing", "verdict": "REJECT" }, { "expected": "REJECT", "name": "held_out_used_for_selection", "verdict": "REJECT" }, { "expected": "REJECT", "name": "held_out_failed", "verdict": "REJECT" }, { "expected": "REJECT", "name": "bad_direction", "verdict": "REJECT" }, { "expected": "REJECT", "name": "duplicate_file", "verdict": "REJECT" }, { "expected": "REJECT", "name": "malformed_file_hash", "verdict": "REJECT" }, { "expected": "REJECT", "name": "artifact_hash_missing", "verdict": "REJECT" }, { "expected": "REJECT", "name": "experiment_not_linked", "verdict": "REJECT" }, { "expected": "REJECT", "name": "claim_overstates_improvement", "verdict": "REJECT" }, { "expected": "REJECT", "name": "wrong_schema", "verdict": "REJECT" }, { "expected": "REJECT", "name": "garbage", "verdict": "REJECT" } ], "summary": "VERIFIED: improvement 0.09 over 95% CI 0.02, one failing-then-passing test, no regression; 22/22 change-failure cases as expected" }, "schema": "cain42.agentic75.findings.v1", "sku": "CAIN-A75-009" }