{
  "schema_version": "0.6",
  "generation_sampled": true,
  "skill": {
    "name": "documentation-and-adrs",
    "version": "0.0.0",
    "content_hash": "b67a9f07ed1057963d3d8ab04bb070ee74d93174eee48b563aa235057005c158",
    "tokens": 2454
  },
  "suite": {
    "format": "agentskills.io/evals",
    "suite_hash": "6cce54e7d9d0fc923eda5499c75fcc68974b3d8714b2a726e2ad1d77e4b5acc0",
    "case_count": 1,
    "canary": "86e8fbf6-9c40-ec04-dacc-03e730017455"
  },
  "run": {
    "model_id": "claude-haiku-4-5-20251001",
    "model_release_date": "2025-10-01",
    "provider": "anthropic",
    "surface": "claude-cli",
    "runner_version": "0.10.1",
    "date_utc": "2026-10-07T22:01:07.808Z",
    "registry": "registered",
    "transcripts": "retained-local",
    "judge": {
      "samples": 3,
      "temperature": null,
      "sampling": "surface-controlled",
      "surface": "claude-cli",
      "model_id": "claude-opus-5",
      "prompt_template_hash": "82586d1e44f84df2b8ce5f18ac777866465cb5ad9686503c93d9776c93c344ed"
    },
    "answered_by": {
      "kind": "model",
      "attested": true,
      "reported_model": "claude-haiku-4-5-20251001",
      "reported_models": [
        "claude-haiku-4-5",
        "claude-opus-5"
      ],
      "isolation": "same-user"
    },
    "pricing_snapshot": {
      "frozen_at": "2026-10-07T22:01:07.808Z",
      "source": "config/models.json",
      "currency": "USD",
      "models": {
        "claude-haiku-4-5-20251001": {
          "input_per_mtok": 1,
          "output_per_mtok": 5,
          "registered": true
        },
        "claude-opus-5": {
          "input_per_mtok": 5,
          "output_per_mtok": 25,
          "registered": true
        }
      },
      "note": "Prices frozen at run time. Every derived cost in this receipt is computed from THIS snapshot, never from the live registry, so the receipt keeps its meaning when registry prices later change."
    }
  },
  "results": {
    "cases": [
      {
        "id": "comment-intent-not-implementation",
        "mode": "with_skill",
        "outcome": "borderline",
        "score": 0.739444,
        "mean": 0.739444,
        "stddev": 0.09229,
        "samples": [
          0.8,
          0.8,
          0.8
        ],
        "generation_hash": "863e44adbe3c4c880184c0721b9ddb74571a62ed1990d2e4222f36df0f1696d6",
        "judge_sample_hashes": [
          "99f2c269e5b194895b5f0c54f08ed074ed30eea4833ab041eb024d373a330ddf",
          "f6a9ec9e6804a56505cc01131546cc436941ed0b9aa28fbf47d52d0a46929e25",
          "03e278c51319673e88b1615e553247f84ef6502b631ccbc2560bc910575706b4"
        ],
        "threshold": 0.7,
        "reason": "Removes all three 'what'-restating comments, the commented-out legacyRate line, and the bare TODO; no intent comment added for the gold-tier discount, so not exemplary.",
        "judge": {
          "model_id": "claude-opus-5",
          "rubric_hash": "b38f01e6671fc0b28a94de550fe70f879134a3168b35e5b1fb53a9a6773478b0"
        },
        "judge_usage": {
          "input_tokens": 44454,
          "output_tokens": 1010,
          "cached_tokens": 32891,
          "wall_ms": 36226
        },
        "generation": {
          "n_planned": 3,
          "n_drawn": 6,
          "n_measured": 6,
          "n_unmeasured": 0,
          "stopping_reason": "stabilised",
          "mean": 0.739444,
          "sd": 0.09229,
          "judge_sd_mean": 0.00866,
          "variance_ratio": 10.657044,
          "variance_ratio_unavailable": null,
          "n_truncated": 0,
          "draws": [
            {
              "draw_index": 0,
              "generation_hash": "645d1b5d13864182bdc12799ada4ee27756a5bf75315eca29c358e1a76b259a9",
              "status": "measured",
              "samples": [
                0.8,
                0.8,
                0.8
              ],
              "judge_sample_hashes": [
                "b6ab6ba6388e0da0f3eb1175d3b1cbe9b2b745a26a8513d62d9ec7eea5347bf4",
                "38587d4e154e70baf263646e64a97b808016b4b4ce63d16206894966bb49625d",
                "3b13f9909ea3ea96e7191179b2e9b6bf2b79a05c0d907cbe86153971dacfa058"
              ],
              "mean": 0.8,
              "stddev": 0,
              "stop_reason": "end_turn",
              "truncated": false,
              "reported_model": "claude-haiku-4-5-20251001",
              "usage": {
                "input_tokens": 16600,
                "output_tokens": 879,
                "cached_tokens": 16590,
                "wall_ms": 10624
              },
              "judge_usage": {
                "input_tokens": 44511,
                "output_tokens": 1099,
                "cached_tokens": 32929,
                "wall_ms": 42636
              }
            },
            {
              "draw_index": 1,
              "generation_hash": "2dce94d47322af9e7aa316dbeef7021b5aac1f21749009a192d3050c254c94ea",
              "status": "measured",
              "samples": [
                0.68,
                0.6,
                0.65
              ],
              "judge_sample_hashes": [
                "0f639665c3ccc96a363b22b689ace1886c7b7e51416cf9b552c9939f64b73804",
                "a46fe97b820e10a7c6203bfe56a54b69af73138706a19a8ff13f3e20183aa464",
                "ae46a67077349182091457893be2c27788db71bfbd9ceb8b58058d769c01726f"
              ],
              "mean": 0.643333,
              "stddev": 0.040415,
              "stop_reason": "end_turn",
              "truncated": false,
              "reported_model": "claude-haiku-4-5-20251001",
              "usage": {
                "input_tokens": 16600,
                "output_tokens": 1672,
                "cached_tokens": 16590,
                "wall_ms": 17757
              },
              "judge_usage": {
                "input_tokens": 44682,
                "output_tokens": 1616,
                "cached_tokens": 33043,
                "wall_ms": 29802
              }
            },
            {
              "draw_index": 2,
              "generation_hash": "bf1b53eba77d53260dbf3a6a63c7919e2851a689a98200721807e458ba843aa0",
              "status": "measured",
              "samples": [
                0.8,
                0.8,
                0.8
              ],
              "judge_sample_hashes": [
                "2815aa7f32704b0837981fc37ad542d1aad723282bbfef7fb31f5fb06ff4d442",
                "1aee6c4189e12073fa8f7647f93c48fc37471498ad7c57b24762bfba4fd44285",
                "996acaa9a409b7224da4a86ef5f993299854936750b72b2ca6548f27818cfd68"
              ],
              "mean": 0.8,
              "stddev": 0,
              "stop_reason": "end_turn",
              "truncated": false,
              "reported_model": "claude-haiku-4-5-20251001",
              "usage": {
                "input_tokens": 16600,
                "output_tokens": 1016,
                "cached_tokens": 16590,
                "wall_ms": 12139
              },
              "judge_usage": {
                "input_tokens": 44454,
                "output_tokens": 1201,
                "cached_tokens": 32891,
                "wall_ms": 23419
              }
            },
            {
              "draw_index": 3,
              "generation_hash": "4da40d021b42c7df21b8509ae254aba609f237ea05ba60ad2b29925016c4ed27",
              "status": "measured",
              "samples": [
                0.8,
                0.78,
                0.8
              ],
              "judge_sample_hashes": [
                "40e85d591d73e051c03b7886ec0ab8f9f06a23b3c8905ef3a5982ca68224351a",
                "35453fa13e1ffc6742874eb40a6f6bc8f0df1e093f79b23089b384d5a38e8a59",
                "adcfb108008d5b34e6a18345aa1b3f1e38a9723a64d88fa871f539f99a96be70"
              ],
              "mean": 0.793333,
              "stddev": 0.011547,
              "stop_reason": "end_turn",
              "truncated": false,
              "reported_model": "claude-haiku-4-5-20251001",
              "usage": {
                "input_tokens": 16600,
                "output_tokens": 996,
                "cached_tokens": 16590,
                "wall_ms": 12076
              },
              "judge_usage": {
                "input_tokens": 44448,
                "output_tokens": 1150,
                "cached_tokens": 32887,
                "wall_ms": 44588
              }
            },
            {
              "draw_index": 4,
              "generation_hash": "f35fce3f97da2faa710339ab47b0c166b5134feb248b0b27069f0bcf74561d30",
              "status": "measured",
              "samples": [
                0.6,
                0.6,
                0.6
              ],
              "judge_sample_hashes": [
                "a23c6ec8eeb87443ef02c02062570a94963a1f0d1025e958dce5d6218cfeac98",
                "7b386466bba9b4d107fa1375ee8ff399e135f60b8d9e6818dec2fad4def26b6f",
                "f5c963c03f1860cfca45beaee4e1bb70d671d9391d436dfc79937f60939f7beb"
              ],
              "mean": 0.6,
              "stddev": 0,
              "stop_reason": "end_turn",
              "truncated": false,
              "reported_model": "claude-haiku-4-5-20251001",
              "usage": {
                "input_tokens": 16600,
                "output_tokens": 1498,
                "cached_tokens": 16590,
                "wall_ms": 14675
              },
              "judge_usage": {
                "input_tokens": 44496,
                "output_tokens": 1465,
                "cached_tokens": 32919,
                "wall_ms": 27839
              }
            },
            {
              "draw_index": 5,
              "generation_hash": "863e44adbe3c4c880184c0721b9ddb74571a62ed1990d2e4222f36df0f1696d6",
              "status": "measured",
              "samples": [
                0.8,
                0.8,
                0.8
              ],
              "judge_sample_hashes": [
                "99f2c269e5b194895b5f0c54f08ed074ed30eea4833ab041eb024d373a330ddf",
                "f6a9ec9e6804a56505cc01131546cc436941ed0b9aa28fbf47d52d0a46929e25",
                "03e278c51319673e88b1615e553247f84ef6502b631ccbc2560bc910575706b4"
              ],
              "mean": 0.8,
              "stddev": 0,
              "stop_reason": "end_turn",
              "truncated": false,
              "reported_model": "claude-haiku-4-5-20251001",
              "usage": {
                "input_tokens": 16600,
                "output_tokens": 1156,
                "cached_tokens": 16590,
                "wall_ms": 13812
              },
              "judge_usage": {
                "input_tokens": 44454,
                "output_tokens": 1010,
                "cached_tokens": 32891,
                "wall_ms": 36226
              }
            }
          ]
        }
      },
      {
        "id": "comment-intent-not-implementation",
        "mode": "baseline",
        "outcome": "fail",
        "score": 0.605238,
        "mean": 0.605238,
        "stddev": 0.084258,
        "samples": [
          0.6,
          0.6,
          0.6
        ],
        "generation_hash": "f62afe83c3946bd37de5891b405746b279b7031ec4725c02008039fd575152d7",
        "judge_sample_hashes": [
          "50ad9dea48e6f976b1ca94459576db794a6cc2c197c619c25d78da01841900aa",
          "90c1da50e5c4aaae91eea88797d098eb085b462cc96cc2b63fa93967bb067537",
          "8fc1df6d5f1df2c3c5c7721476821cc404266fb0ab2bf4b25273520fcc7698e1"
        ],
        "threshold": 0.7,
        "reason": "Met (a) removed all 'what' comments, (b) removed commented-out legacy line, (d) no restating comments added; missed (c) left bare TODO dangling and added no intent comment.",
        "judge": {
          "model_id": "claude-opus-5",
          "rubric_hash": "b38f01e6671fc0b28a94de550fe70f879134a3168b35e5b1fb53a9a6773478b0"
        },
        "judge_usage": {
          "input_tokens": 44544,
          "output_tokens": 879,
          "cached_tokens": 32951,
          "wall_ms": 31339
        },
        "generation": {
          "n_planned": 3,
          "n_drawn": 7,
          "n_measured": 7,
          "n_unmeasured": 0,
          "stopping_reason": "stabilised",
          "mean": 0.605238,
          "sd": 0.084258,
          "judge_sd_mean": 0.010722,
          "variance_ratio": 7.858422,
          "variance_ratio_unavailable": null,
          "n_truncated": 0,
          "draws": [
            {
              "draw_index": 0,
              "generation_hash": "184a1a50b4aa5b5a3e4bb823599d6a90af9ae60c69c08cb240ffc7d958aabff0",
              "status": "measured",
              "samples": [
                0.6,
                0.6,
                0.6
              ],
              "judge_sample_hashes": [
                "eb05a87dbcb755d5a5d51648337359b2338977013f5250933f2b9bf9fcd0be16",
                "6f6144a95de238d047dc2bdf691d94fe222c78d3231b5ddfaceeda91ed74fa0a",
                "ac8b9a616dea8ff9a8f75b653e40070d19519c982556667d0be10cb325e3281e"
              ],
              "mean": 0.6,
              "stddev": 0,
              "stop_reason": "end_turn",
              "truncated": false,
              "reported_model": "claude-haiku-4-5-20251001",
              "usage": {
                "input_tokens": 14139,
                "output_tokens": 841,
                "cached_tokens": 14129,
                "wall_ms": 10321
              },
              "judge_usage": {
                "input_tokens": 44397,
                "output_tokens": 1258,
                "cached_tokens": 32853,
                "wall_ms": 24435
              }
            },
            {
              "draw_index": 1,
              "generation_hash": "c7c0dcf2ec3772b5db403862b4b24893404aad9a3851dff009b60098f29242fe",
              "status": "measured",
              "samples": [
                0.5,
                0.5,
                0.45
              ],
              "judge_sample_hashes": [
                "ea098fb7c6ed6523f4f37829021756ca30e2a15fb0679e46d52264348ffa7f0d",
                "05d22409035674d754490a68862a567a56100064eda20ce5839f2b25f019c06c",
                "b64a23b214a86cffd4bd8dffd37b11c6e2203f09b76daa97e17c7ae8353698f3"
              ],
              "mean": 0.483333,
              "stddev": 0.028868,
              "stop_reason": "end_turn",
              "truncated": false,
              "reported_model": "claude-haiku-4-5-20251001",
              "usage": {
                "input_tokens": 14139,
                "output_tokens": 918,
                "cached_tokens": 14129,
                "wall_ms": 11837
              },
              "judge_usage": {
                "input_tokens": 44541,
                "output_tokens": 1891,
                "cached_tokens": 32949,
                "wall_ms": 33512
              }
            },
            {
              "draw_index": 2,
              "generation_hash": "714c65f56d5644f26005d7d5cdf128a4676e07f2a1304a6002154c348f0d3550",
              "status": "measured",
              "samples": [
                0.6,
                0.6,
                0.6
              ],
              "judge_sample_hashes": [
                "015f620f182f284ceea79f0b090170b565619254e91c069703d23ca088810147",
                "3b5435008944759b9c765814b65fedc07deb88363576e25e4800a66e13489e04",
                "44df27e3bb9248ee5087b30a3d2d3a5bcf67dd4f31652fe7e3955265dfc1c665"
              ],
              "mean": 0.6,
              "stddev": 0,
              "stop_reason": "end_turn",
              "truncated": false,
              "reported_model": "claude-haiku-4-5-20251001",
              "usage": {
                "input_tokens": 14139,
                "output_tokens": 1627,
                "cached_tokens": 14129,
                "wall_ms": 17212
              },
              "judge_usage": {
                "input_tokens": 44436,
                "output_tokens": 1316,
                "cached_tokens": 32879,
                "wall_ms": 27112
              }
            },
            {
              "draw_index": 3,
              "generation_hash": "8322f68a59e74bbeb005ff177cef4cae3c69b24febda12da29b06577931ac803",
              "status": "measured",
              "samples": [
                0.78,
                0.78,
                0.75
              ],
              "judge_sample_hashes": [
                "6ba6329bfdd81854532f6391685c821040e315bd879ad724967089feac0ba3c4",
                "ac1f1449c02f902c514ecc61274fe57f0fb08c6500e1b00dd745f462b36de9f1",
                "21121f9bd364af46c6c453fbff2d636877bf43bbd91d7d1b2c904aba9e95916b"
              ],
              "mean": 0.77,
              "stddev": 0.017321,
              "stop_reason": "end_turn",
              "truncated": false,
              "reported_model": "claude-haiku-4-5-20251001",
              "usage": {
                "input_tokens": 14139,
                "output_tokens": 1237,
                "cached_tokens": 14129,
                "wall_ms": 14098
              },
              "judge_usage": {
                "input_tokens": 44769,
                "output_tokens": 1571,
                "cached_tokens": 33101,
                "wall_ms": 41663
              }
            },
            {
              "draw_index": 4,
              "generation_hash": "6ea5d004e8467119c6791b241ab1fda4ca6d9bcd5282ebeda32d16e5329c4f4b",
              "status": "measured",
              "samples": [
                0.6,
                0.6,
                0.55
              ],
              "judge_sample_hashes": [
                "8951f6b3ae5996c5b45ef889a8968d8d743ebeb10c25933b0d161997e4f42d32",
                "5079321fe1de76f69ea294d9e7e4bc6870882c9898156807958ac92d86e953b0",
                "37d81118a429f35bb7bbf6a35f39b1bfc929d204bd10d5b288cf97d9631c1542"
              ],
              "mean": 0.583333,
              "stddev": 0.028868,
              "stop_reason": "end_turn",
              "truncated": false,
              "reported_model": "claude-haiku-4-5-20251001",
              "usage": {
                "input_tokens": 14139,
                "output_tokens": 1024,
                "cached_tokens": 14129,
                "wall_ms": 11855
              },
              "judge_usage": {
                "input_tokens": 44565,
                "output_tokens": 1313,
                "cached_tokens": 32965,
                "wall_ms": 78514
              }
            },
            {
              "draw_index": 5,
              "generation_hash": "d75ebd159643d2576dd3507604c8ceeca85c76e1bca83009ec144b429e1c8401",
              "status": "measured",
              "samples": [
                0.6,
                0.6,
                0.6
              ],
              "judge_sample_hashes": [
                "2b51f745317e276737300de2d337d7e8bf89fd1ecb432204875f1a58a33b09a7",
                "f32faa947578829acc4a835cc09dc347c016937e1e517133a24b305d6e7a2af2",
                "e7449a60db8c9ce20023a0d566dc9a96a885f7642115277a23b9259fe13a6adb"
              ],
              "mean": 0.6,
              "stddev": 0,
              "stop_reason": "end_turn",
              "truncated": false,
              "reported_model": "claude-haiku-4-5-20251001",
              "usage": {
                "input_tokens": 14139,
                "output_tokens": 1021,
                "cached_tokens": 14129,
                "wall_ms": 12265
              },
              "judge_usage": {
                "input_tokens": 44529,
                "output_tokens": 916,
                "cached_tokens": 32941,
                "wall_ms": 20782
              }
            },
            {
              "draw_index": 6,
              "generation_hash": "f62afe83c3946bd37de5891b405746b279b7031ec4725c02008039fd575152d7",
              "status": "measured",
              "samples": [
                0.6,
                0.6,
                0.6
              ],
              "judge_sample_hashes": [
                "50ad9dea48e6f976b1ca94459576db794a6cc2c197c619c25d78da01841900aa",
                "90c1da50e5c4aaae91eea88797d098eb085b462cc96cc2b63fa93967bb067537",
                "8fc1df6d5f1df2c3c5c7721476821cc404266fb0ab2bf4b25273520fcc7698e1"
              ],
              "mean": 0.6,
              "stddev": 0,
              "stop_reason": "end_turn",
              "truncated": false,
              "reported_model": "claude-haiku-4-5-20251001",
              "usage": {
                "input_tokens": 14139,
                "output_tokens": 859,
                "cached_tokens": 14129,
                "wall_ms": 9826
              },
              "judge_usage": {
                "input_tokens": 44544,
                "output_tokens": 879,
                "cached_tokens": 32951,
                "wall_ms": 31339
              }
            }
          ]
        }
      }
    ],
    "aggregates": {
      "with_skill": {
        "case_count": 1,
        "pass_count": 0,
        "borderline_count": 1,
        "mean_score": 0.739444,
        "stddev": null
      },
      "baseline": {
        "case_count": 1,
        "pass_count": 0,
        "borderline_count": 0,
        "mean_score": 0.605238,
        "stddev": null
      },
      "band_rule": "per arm: the sample standard deviation (n-1) of the per-case means across the included cases; the comparison band is the quadrature sum of the two arms; null where an arm has fewer than two included cases"
    }
  },
  "comparison": {
    "with_skill_score": 0.739444,
    "baseline_score": 0.605238,
    "delta": 0.134206,
    "delta_uncertainty": null,
    "delta_uncertainty_unavailable": "single_case"
  },
  "verification_level": "TESTED",
  "receipt_hash": "c298235bf3e35b729a3024ae1d656af4e7120fcec1c16bfffc8eb3853fb9c456",
  "economics": {
    "basis": "metered-equivalent",
    "surface": "claude-cli",
    "with_skill": {
      "call_count": 6,
      "mean_input_tokens": 16600,
      "mean_output_tokens": 1202.83,
      "mean_cost_usd_per_call": 0.022614,
      "median_wall_ms": 12975.5,
      "wall_ms_p25": 12076,
      "wall_ms_p75": 14675,
      "wall_ms_iqr": 2599
    },
    "baseline": {
      "call_count": 7,
      "mean_input_tokens": 14139,
      "mean_output_tokens": 1075.29,
      "mean_cost_usd_per_call": 0.019515,
      "median_wall_ms": 11855,
      "wall_ms_p25": 10321,
      "wall_ms_p75": 14098,
      "wall_ms_iqr": 3777
    },
    "skill_incremental_cost_usd_per_call": 0.003099,
    "skill_incremental_cost_usd_per_1k_calls": 3.099,
    "output_tokens_delta": 127.54,
    "median_wall_ms_delta": 1120.5,
    "judge_excluded": true,
    "judge_overhead": {
      "note": "Measurement overhead imposed by Driftproof, NOT a cost of running the skill. Excluded from every field above.",
      "total_cost_usd": 0.492215,
      "case_rows_measured": 2
    },
    "notes": {
      "absolute_cost": "Absolute per-call cost on a CLI surface includes a fixed harness preamble (observed ~25k input tokens on claude-cli, ~11k on codex) that we do not control. It is identical in the with-skill and baseline arms, so it cancels in the Δ figures — read the incremental columns, not the absolute ones.",
      "cache_pricing": "Cached input tokens are costed at the list input rate (an over-estimate where caching is heavy); cached_tokens is recorded per call so a reader can recompute.",
      "latency": "observed on subscription CLI surface, indicative"
    }
  }
}