{
  "anime": {
    "schema_version": "codevetter-autonomous-browser-loop-proof/v1",
    "recorded_at": "2026-08-14",
    "claim": "On an isolated Anime List checkout, CodeVetter rejected sub-threshold candidates and retained one two-file local-development loading improvement after correctness-first screening and ten-sample paired promotion.",
    "product": "anime-list",
    "subject": {
      "repository_revision": "db1faf0e0c27996c3eacfb9f1e3a4c0a65958938",
      "baseline_source_snapshot_sha256": "01de0824927741ce93da7ff9eb4334d772711d451ec96ff166f5874d908db19d",
      "retained_source_snapshot_sha256": "c3a862e3d50f322b4d201c7e10c970fc3d6fcabc8a3dd475039c64709c386145",
      "isolation": "disposable detached worktrees with copied package payload and an empty HOME",
      "network_or_cloud": false
    },
    "flow": {
      "target": "e2e/mobile.spec.ts",
      "name": "home page has no horizontal scroll",
      "browser_project": "mobile",
      "scope": "one exact local Chromium navigation against CodeVetter's config-disabled owned Vite runtime"
    },
    "first_pass": {
      "observed_families": [
        "browser_timing",
        "memory",
        "react",
        "actions",
        "dependencies"
      ],
      "incomplete_families": [
        "loading"
      ],
      "unavailable_families": [
        "build_artifact",
        "review"
      ],
      "cause_groups": 6,
      "queued_experiments": 4,
      "highest_measured_dependency": {
        "package": "lucide-react",
        "development_response_bytes": 4263394,
        "exact_initial_importers": [
          "components/Navigation.tsx",
          "components/ui/dropdown-menu.tsx"
        ]
      }
    },
    "rejected_experiments": [
      {
        "hypothesis": "Replace five radix-ui package-root imports with package subpaths.",
        "changed_files": 5,
        "screening_samples_per_side": 3,
        "baseline_completed_response_bytes_median": 12979530,
        "candidate_completed_response_bytes_median": 11778516,
        "delta_bytes": -1201014,
        "delta_percent": -9.254,
        "decision": "rejected",
        "reason": "The measured movement did not clear the 10 percent transfer materiality floor.",
        "loop_elapsed_minutes_at_decision": 0.788
      },
      {
        "hypothesis": "Memoize HomePage to remove repeated React work.",
        "changed_files": 1,
        "screening_samples_per_side": 3,
        "decision": "rejected",
        "reason": "No exact-flow timing, memory, React, or loading metric cleared materiality."
      }
    ],
    "retained_experiment": {
      "experiment_id": "bcbcd81274431f6792a3ef74",
      "hypothesis": "Remove lucide-react from the exact initial home-flow boundary while preserving deferred-route usage.",
      "host_patch": "Inline four small SVG icons in the two statically reachable importers; deferred routes remain unchanged.",
      "changed_files": [
        "components/Navigation.tsx",
        "components/ui/dropdown-menu.tsx"
      ],
      "complexity": {
        "files_changed": 2,
        "added_lines": 60,
        "deleted_lines": 2,
        "gross_lines_changed": 62,
        "net_lines_changed": 58,
        "production_dependencies_added": [],
        "budget": {
          "max_files_changed": 3,
          "max_lines_added": 160,
          "max_gross_lines_changed": 200,
          "max_production_dependencies_added": 0
        },
        "violations": []
      },
      "screening": {
        "samples_per_side": 3,
        "browser_completed_response_transfer_bytes": {
          "baseline_median": 11871129,
          "candidate_median": 7616611,
          "delta_bytes": -4254518,
          "delta_percent": -35.839
        },
        "decision": "promote"
      },
      "promotion": {
        "samples_per_side": 10,
        "warmups_per_side": 1,
        "browser_completed_response_transfer_bytes": {
          "baseline_median": 11871129,
          "candidate_median": 7616611,
          "delta_bytes": -4254518,
          "delta_percent": -35.839,
          "stable": true
        },
        "workload_duration_ms": {
          "baseline_median": 316.5115,
          "candidate_median": 303.9395,
          "delta_percent": -3.972,
          "material": false,
          "regressed": false
        },
        "largest_contentful_paint_ms": {
          "baseline_median": 250.916,
          "candidate_median": 236.1995,
          "delta_percent": -5.865,
          "material": false,
          "regressed": false
        },
        "process_tree_peak_rss_bytes": {
          "baseline_median": 861634560,
          "candidate_median": 819912704,
          "delta_percent": -4.842,
          "material": false,
          "regressed": false
        },
        "react_actual_duration_ms": {
          "baseline_median": 17.55,
          "candidate_median": 17.2,
          "delta_percent": -1.994,
          "material": false,
          "regressed": false
        },
        "correctness": {
          "adapter": "vitest",
          "target": "src/recommendations.test.ts",
          "name": "keeps only the highest-ranked recommendations in stable order",
          "incumbent_selected": 1,
          "incumbent_failed": 0,
          "candidate_selected": 1,
          "candidate_failed": 0
        },
        "decision": "kept",
        "reason": "Correctness passed and stable paired evidence met the promotion policy."
      },
      "campaign_elapsed_minutes_from_initialization_to_keep": 2.315,
      "next_generation_replanned": true
    },
    "production_build_check": {
      "provenance": "Separate host-run Vite production builds with an empty HOME; this is a diagnostic qualification, not the loop's acceptance metric.",
      "baseline_initial_javascript": {
        "raw_bytes": 504556,
        "gzip_bytes": 159923
      },
      "candidate_initial_javascript": {
        "raw_bytes": 503474,
        "gzip_bytes": 159423
      },
      "delta": {
        "raw_bytes": -1082,
        "raw_percent": -0.214,
        "gzip_bytes": -500,
        "gzip_percent": -0.313
      },
      "conclusion": "The retained result is a material local Vite development-navigation improvement, not a material shipped initial-bundle improvement."
    },
    "superseded_measurement": {
      "previous_claimed_gzip_change_percent": -28.452,
      "reason": "That comparison measured selected chunk redistribution rather than the complete HTML-referenced initial JavaScript closure.",
      "correct_full_closure_chunk_rule_result": {
        "baseline_gzip_bytes": 159923,
        "candidate_gzip_bytes": 154162,
        "delta_bytes": -5761,
        "delta_percent": -3.602,
        "decision": "rejected_below_materiality"
      }
    },
    "manual_vs_loop": {
      "prior_manual_investigation": {
        "agent_actions": [
          "inspect source and historical diff",
          "construct four build variants",
          "manually compare chunk output",
          "run a separate non-interleaved browser replay"
        ],
        "outcome": "Mechanism hypothesis only; no candidate retained, and the headline percentage was later found to use an incomplete closure."
      },
      "autonomous_loop": {
        "host_actions_per_experiment": [
          "apply the returned bounded edit",
          "call evaluate-browser-experiment",
          "restore only when rejected"
        ],
        "codevetter_owned": [
          "evidence collection and dependency ranking",
          "exact static-versus-deferred source seal",
          "correctness selection",
          "alternating screening captures",
          "automatic ten-sample promotion",
          "materiality and regression gates",
          "append-only keep or reject decision",
          "incumbent advancement and replanning"
        ],
        "outcome": "One sub-threshold candidate rejected and one local-development loading candidate retained with repeated evidence."
      }
    },
    "product_improvements_discovered_during_trial": [
      "Dirty candidates are accepted only when every changed file is inside the exact sealed experiment boundary and the snapshot is rechecked around every capture.",
      "Multiline static imports are included in the literal graph.",
      "Initial and deferred importers are retained separately, preventing lazy-route files from inflating the edit seal.",
      "Large captured Vite dependency routes join to all exact static importers rather than substring-matching unrelated packages.",
      "Transfer verification prefers the complete zero-failure navigation action cohort, excluding later evaluation and analytics request noise."
    ],
    "limitations": [
      "The keep proves one local config-disabled Vite development navigation, not production traffic, representative devices, CDN behavior, or end-user impact.",
      "The correctness scope is one exact repository-owned recommendation test plus the selected Playwright flow; it is not the full Anime List suite.",
      "The production build check shows only a 0.313 percent gzip reduction, so this candidate should be presented as local developer-flow optimization rather than a consumer performance case study.",
      "The host agent still invents and applies source patches; CodeVetter owns evidence, boundaries, verification, decisions, and replanning.",
      "No production, cloud, registry, package installation, migration, deployment, paid-model, or secret operation was used."
    ]
  },
  "free_ai": {
    "schema_version": "codevetter-local-optimization-proof/v1",
    "recorded_at": "2026-08-14T16:47:50+05:30",
    "observation_date": "2026-08-14",
    "product": "free-ai",
    "claim": "On one isolated local Free AI checkout, a one-file candidate removed the intermediate eligibility array from model selection. Ten interleaved samples confirmed 29.841% lower largest-input time and 30.139% lower sampled bytes at the selected allocation source, with no material peak-RSS regression.",
    "subject": {
      "repository_revision": "23ed8566bb1b1d8d86ee1f9890aa4ad5bba9cd18",
      "baseline_source_snapshot_sha256": "93b01ba755f1e22c53180e5c21b232b806e3d2302f7b11fc72e71a12c49542f4",
      "candidate_source_snapshot_sha256": "d30851f91295807b466f6319bb467114225944acf794a8b39b4fe45751e337ce",
      "isolation": "detached local Git worktree with a copy-on-write dependency tree",
      "network_or_cloud": false
    },
    "flow": {
      "adapter": "vitest",
      "target": "test/select-model-performance.spec.ts",
      "name": "model selection scales across the supported registry size",
      "largest_input": 79,
      "iterations_per_input": 5000,
      "workload_identity": {
        "algorithm": "sha256",
        "digest": "08441b9854d61439d7033983fff0aa4526a3ac74f87336993a3d05b7ec661629"
      }
    },
    "diagnosis": {
      "kind": "application_allocation_hotspot",
      "source": "src/router/select-model.ts:171",
      "function": "selectCandidates",
      "initial_sampled_allocation_share": 0.841449,
      "hypothesis": "Build the ranked candidate list directly instead of allocating an intermediate filtered array."
    },
    "change_cost": {
      "files_changed": 1,
      "changed_files": [
        "src/router/select-model.ts"
      ],
      "lines_added": 19,
      "lines_removed": 26,
      "gross_lines_changed": 45,
      "net_lines_changed": -7,
      "production_dependencies_added": [],
      "policy": {
        "max_files_changed": 3,
        "max_lines_added": 160,
        "max_gross_lines_changed": 200,
        "max_production_dependencies_added": 0
      },
      "violations": []
    },
    "paired_verification": {
      "evidence_mode": "paired_interleaved",
      "samples_per_side": 10,
      "warmups_per_side": 1,
      "scale_points": [
        {
          "input": 20,
          "baseline_ms_per_operation": 0.006524,
          "candidate_ms_per_operation": 0.004648,
          "delta_percent": -28.755
        },
        {
          "input": 50,
          "baseline_ms_per_operation": 0.008153,
          "candidate_ms_per_operation": 0.005679,
          "delta_percent": -30.345
        },
        {
          "input": 79,
          "baseline_ms_per_operation": 0.016568,
          "candidate_ms_per_operation": 0.011624,
          "delta_percent": -29.841
        }
      ],
      "sampled_source_bytes": {
        "baseline_median": 220208512,
        "candidate_median": 153840808,
        "delta_bytes": -66367704,
        "delta_percent": -30.139,
        "attribution_status": "source_and_application_total_agree"
      },
      "peak_process_tree_rss_bytes": {
        "baseline_median": 227721216,
        "candidate_median": 227639296,
        "delta_bytes": -81920,
        "delta_percent": -0.036,
        "material": false,
        "regressed": false
      },
      "decision": {
        "mechanically_confirmed": true,
        "materially_useful": true,
        "shipping_recommended": true,
        "verdict": "confirmed",
        "reason": "Largest-input time improved by 29.841% without a material smaller-input regression."
      }
    },
    "correctness": {
      "focused_selection_tests": 27,
      "full_test_files_passed": 37,
      "full_tests_passed": 241,
      "typecheck": "passed",
      "lint_changed_file": "passed"
    },
    "limitations": [
      "This proves one synthetic local model-registry workload up to the repository's current 79-model registry; it does not establish production request latency or customer impact.",
      "Sampled heap bytes include allocations that may later be collected and are not retained-heap measurements.",
      "Peak RSS is sampled process-tree evidence and includes the runtime and test runner.",
      "No production endpoint, provider API, cloud resource, package installation, deployment, paid model, or secret operation was used."
    ]
  },
  "starboard": {
    "schema_version": "codevetter-local-optimization-rejection/v1",
    "recorded_at": "2026-08-14",
    "product": "starboard",
    "repository_revision": "7af23ce31ac10c7c552713f92985934de116abee",
    "flow": {
      "adapter": "vitest",
      "target": "src/__tests__/project-recommendations-performance.test.ts",
      "name": "project recommendations scale across local catalog sizes",
      "workload_identity_sha256": "7ea08cb2f0d112220eaedabd2103336419473773a23a3a774b1b98c4c4abb6ba"
    },
    "diagnosis": {
      "source": "src/lib/project-recommendations.ts:101",
      "function": "meaningfulTokens",
      "cpu_sample_share": 0.1084
    },
    "experiment": {
      "summary": "Replace join, match, filter, and Set construction with an incremental regular-expression scan.",
      "change_cost": {
        "files_changed": 1,
        "lines_added": 12,
        "lines_removed": 7,
        "gross_lines_changed": 19,
        "production_dependencies_added": [],
        "budget_violations": []
      },
      "focused_correctness": {
        "test_files_passed": 2,
        "tests_passed": 13
      }
    },
    "paired_verification": {
      "samples_per_side": 10,
      "warmups_per_side": 1,
      "largest_input": 50000,
      "baseline_ms_per_operation": 47.496,
      "candidate_ms_per_operation": 50.026,
      "delta_percent": 5.327,
      "peak_rss_delta_percent": -0.55,
      "allocation_evidence": "incomplete_collection_bound_exceeded",
      "verdict": "no_confidence",
      "shipping_recommended": false
    },
    "decision": "rejected_and_restored",
    "limitations": [
      "Wall-time sample spreads were 86.332% and 91.255%, so host or startup noise may dominate.",
      "The candidate was slower at the largest input and heap evidence was incomplete; no optimization claim is made.",
      "No production, cloud, deployment, package installation, paid model, or secret operation was used."
    ]
  }
}