{
  "schema_version": "0.7",
  "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-opus-5-5",
    "model_release_date": null,
    "provider": "anthropic",
    "surface": "claude-cli",
    "runner_version": "0.11.2",
    "date_utc": "2026-09-24T07:18:11.322Z",
    "registry": "registered",
    "transcripts": "hashes-only",
    "judge": {
      "samples": 3,
      "temperature": null,
      "sampling": "surface-controlled",
      "surface": "claude-cli",
      "model_id": "claude-opus-5-5",
      "prompt_template_hash": "82586d1e44f84df2b8ce5f18ac777866465cb5ad9686503c93d9776c93c344ed"
    },
    "answered_by": {
      "kind": "model",
      "attested": true,
      "reported_model": "claude-opus-5-5",
      "reported_models": [
        "claude-opus-5",
        "claude-opus-5-5"
      ],
      "isolation": "same-user"
    },
    "pricing_snapshot": {
      "frozen_at": "2026-09-24T07:18:11.322Z",
      "source": "config/models.json",
      "currency": "USD",
      "models": {
        "claude-opus-5-5": {
          "input_per_mtok": 4,
          "output_per_mtok": 20,
          "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": 0.657778,
        "mean": 0.657778,
        "stddev": 0.105043,
        "samples": [
          0.6,
          0.6,
          0.6
        ],
        "generation_hash": "f4ed0167d112d4c095315111c9764efddc5e05344aa2e86d52dd4b8cc1f723e2",
        "judge_sample_hashes": [
          "6ddaffe6842f93a87a16ca3ad894e3c9bd73f9222c3e4c3e291a4a0f3f7ef1f4",
          "cde6be9d1387282dd2fa0df6ba6d4feebb8dd593ef5024c2b04e063749b9c560",
          "c706e6bad9a367279b3487722e08ac73612331818daf49fdc21965894ad71f67"
        ],
        "threshold": 0.7,
        "reason": "Meets (a), (b), and adds a useful side-effect doc comment. Misses (c): the TODO remains with a placeholder reason. The '10% gold-tier discount' comment only partly states intent (d).",
        "judge": {
          "model_id": "claude-opus-5-5",
          "rubric_hash": "b38f01e6671fc0b28a94de550fe70f879134a3168b35e5b1fb53a9a6773478b0"
        },
        "judge_usage": {
          "input_tokens": 15918,
          "output_tokens": 1171,
          "cached_tokens": 12859,
          "wall_ms": 21425
        },
        "generation": {
          "n_planned": 3,
          "n_drawn": 6,
          "n_measured": 6,
          "n_unmeasured": 0,
          "stopping_reason": "stabilised",
          "mean": 0.657778,
          "sd": 0.105043,
          "judge_sd_mean": 0.046164,
          "variance_ratio": 2.275431,
          "variance_ratio_unavailable": null,
          "n_truncated": 0,
          "draws": [
            {
              "draw_index": 0,
              "generation_hash": "c7276cb74f981e1cc536279f25197b31c4aefb2fdc0dd917e187f8ebff785457",
              "status": "measured",
              "samples": [
                0.55,
                0.4,
                0.55
              ],
              "judge_sample_hashes": [
                "8d1a83469958437555805829217e2036ae69398cc634dc198aa3891bc6e45a2d",
                "5c8797f78324839adc88c64771647a6c464c80bbdb4d73f253a856a83952439c",
                "1b99abcf0db7b15b01780814ccdae43d75e4b320f885941e44153c88b51ea87a"
              ],
              "mean": 0.5,
              "stddev": 0.086603,
              "stop_reason": "end_turn",
              "truncated": false,
              "reported_model": "claude-opus-5-5",
              "usage": {
                "input_tokens": 7187,
                "output_tokens": 840,
                "cached_tokens": 540,
                "wall_ms": 10773
              },
              "judge_usage": {
                "input_tokens": 15816,
                "output_tokens": 833,
                "cached_tokens": 12791,
                "wall_ms": 17835
              }
            },
            {
              "draw_index": 1,
              "generation_hash": "283e07542179f8f516e1486ff00010a649c9a06f8024bd1c26b819605dab2901",
              "status": "measured",
              "samples": [
                0.8,
                0.8,
                0.7
              ],
              "judge_sample_hashes": [
                "fd0e8d00fe93af84c1363ed7845d84d60bc3e34812c7fc988830d7623f0ef291",
                "4a1869fc5933427fe016368bf1289fe942a6c448f04baacbd5d56af35f996a58",
                "4c729796958e7dd3d94843a2fb6e016726b0ade35596566981c0c475cf9e8882"
              ],
              "mean": 0.766667,
              "stddev": 0.057735,
              "stop_reason": "end_turn",
              "truncated": false,
              "reported_model": "claude-opus-5-5",
              "usage": {
                "input_tokens": 7187,
                "output_tokens": 808,
                "cached_tokens": 7185,
                "wall_ms": 10607
              },
              "judge_usage": {
                "input_tokens": 15786,
                "output_tokens": 821,
                "cached_tokens": 12771,
                "wall_ms": 17630
              }
            },
            {
              "draw_index": 2,
              "generation_hash": "b2d75929af9576560652cf901b7a7234f8896dca79bd19d041fda0dbcada8bb4",
              "status": "measured",
              "samples": [
                0.62,
                0.6,
                0.6
              ],
              "judge_sample_hashes": [
                "db9cf83db864c8de2d8e780d6399efbcfb981adbb84fe0ba97a3cc36addc0c64",
                "83b3e1bf8c61d78b5892ecd91e61fb4611b722f67f9c90606540895b8cb056cc",
                "0a1f38a927e72a1a512d42a9650f79abebc729e5c477481fe3f7b173bfc51b23"
              ],
              "mean": 0.606667,
              "stddev": 0.011547,
              "stop_reason": "end_turn",
              "truncated": false,
              "reported_model": "claude-opus-5-5",
              "usage": {
                "input_tokens": 7187,
                "output_tokens": 829,
                "cached_tokens": 7185,
                "wall_ms": 10878
              },
              "judge_usage": {
                "input_tokens": 15570,
                "output_tokens": 1086,
                "cached_tokens": 12627,
                "wall_ms": 19973
              }
            },
            {
              "draw_index": 3,
              "generation_hash": "7a9065550fdd632dbd133d985a10df4888dc7736b1a8b01783853d4ddf87b9b1",
              "status": "measured",
              "samples": [
                0.68,
                0.8,
                0.72
              ],
              "judge_sample_hashes": [
                "2a0e761d68ab985eb1220188af6142d6e1d2bdbfbb481ab82353e2a162400e58",
                "3d9e5cbc99dcee264fe3f5a1955cf978285d036784485aeabd41921805982f0f",
                "1622d06241ec56d975892788f602e35636bb9c7d1b89c275289947b662418773"
              ],
              "mean": 0.733333,
              "stddev": 0.061101,
              "stop_reason": "end_turn",
              "truncated": false,
              "reported_model": "claude-opus-5-5",
              "usage": {
                "input_tokens": 7187,
                "output_tokens": 936,
                "cached_tokens": 7185,
                "wall_ms": 12048
              },
              "judge_usage": {
                "input_tokens": 16026,
                "output_tokens": 1097,
                "cached_tokens": 12931,
                "wall_ms": 19612
              }
            },
            {
              "draw_index": 4,
              "generation_hash": "4b5b0738272a3852070841c4da086f8f3e71a08cf8daf7629aa8a2e45101bad6",
              "status": "measured",
              "samples": [
                0.68,
                0.8,
                0.74
              ],
              "judge_sample_hashes": [
                "495dfc0ba60e27f7b7f6292ebf518484d05d62a0c1dbeb6bda841948a2b8382e",
                "308cb9635feb5022b9cc321a79482889b9b1204995ba021ce9d418482e555627",
                "a08ded4d24e6b41c7fe1aaa00629dc240374cf1b7f4e10b2275bda9e3493c8e1"
              ],
              "mean": 0.74,
              "stddev": 0.06,
              "stop_reason": "end_turn",
              "truncated": false,
              "reported_model": "claude-opus-5-5",
              "usage": {
                "input_tokens": 7187,
                "output_tokens": 889,
                "cached_tokens": 7185,
                "wall_ms": 10835
              },
              "judge_usage": {
                "input_tokens": 16065,
                "output_tokens": 980,
                "cached_tokens": 12957,
                "wall_ms": 18642
              }
            },
            {
              "draw_index": 5,
              "generation_hash": "f4ed0167d112d4c095315111c9764efddc5e05344aa2e86d52dd4b8cc1f723e2",
              "status": "measured",
              "samples": [
                0.6,
                0.6,
                0.6
              ],
              "judge_sample_hashes": [
                "6ddaffe6842f93a87a16ca3ad894e3c9bd73f9222c3e4c3e291a4a0f3f7ef1f4",
                "cde6be9d1387282dd2fa0df6ba6d4feebb8dd593ef5024c2b04e063749b9c560",
                "c706e6bad9a367279b3487722e08ac73612331818daf49fdc21965894ad71f67"
              ],
              "mean": 0.6,
              "stddev": 0,
              "stop_reason": "end_turn",
              "truncated": false,
              "reported_model": "claude-opus-5-5",
              "usage": {
                "input_tokens": 7187,
                "output_tokens": 841,
                "cached_tokens": 7185,
                "wall_ms": 10882
              },
              "judge_usage": {
                "input_tokens": 15918,
                "output_tokens": 1171,
                "cached_tokens": 12859,
                "wall_ms": 21425
              }
            }
          ]
        }
      },
      {
        "id": "comment-intent-not-implementation",
        "mode": "baseline",
        "outcome": "fail",
        "score": 0.586667,
        "mean": 0.586667,
        "stddev": 0.012018,
        "samples": [
          0.58,
          0.55,
          0.6
        ],
        "generation_hash": "294a63179dd624e32b3121cb9b74d9eb7d738ee60ddbccc783713c2cb8dad2db",
        "judge_sample_hashes": [
          "9046a910c3a44c9ca9183d0d502504433596472ffdae13566e9a61d9bb72f553",
          "7fe7bc76ee9fde13943fee009cf3c01617335109fbf3d1d31265d3ed2c9b84c3",
          "cc506ae22d6f5f3635b7576e1f0284f914d1b394fd656194693c37512a4fe182"
        ],
        "threshold": 0.7,
        "reason": "Meets (a) and (b); misses (c) by rewording rather than removing the TODO; (d) partial: side-effect comment explains why, but 'Gold tier gets 10% off' restates the code.",
        "judge": {
          "model_id": "claude-opus-5-5",
          "rubric_hash": "b38f01e6671fc0b28a94de550fe70f879134a3168b35e5b1fb53a9a6773478b0"
        },
        "judge_usage": {
          "input_tokens": 15366,
          "output_tokens": 1238,
          "cached_tokens": 12491,
          "wall_ms": 20881
        },
        "generation": {
          "n_planned": 3,
          "n_drawn": 3,
          "n_measured": 3,
          "n_unmeasured": 0,
          "stopping_reason": "min_reached",
          "mean": 0.586667,
          "sd": 0.012018,
          "judge_sd_mean": 0.018011,
          "variance_ratio": 0.667259,
          "variance_ratio_unavailable": null,
          "n_truncated": 0,
          "draws": [
            {
              "draw_index": 0,
              "generation_hash": "c4fa9e2bf62d5df93b2ba7f86688d91002704b7d97726dfb3da799c082e36046",
              "status": "measured",
              "samples": [
                0.6,
                0.6,
                0.55
              ],
              "judge_sample_hashes": [
                "137cb592b58a385ce6b614eddf75ff1479b5af0088fe45697f80bbac115175a9",
                "b212db87183e8bf52c954323f8b5294e2d4f27db018f751f89029f794813504c",
                "ad03dcc2310d80b2de4ef3e990b9310b6a9a0155ee7d7a3c6796fa6b676bf13d"
              ],
              "mean": 0.583333,
              "stddev": 0.028868,
              "stop_reason": "end_turn",
              "truncated": false,
              "reported_model": "claude-opus-5-5",
              "usage": {
                "input_tokens": 3724,
                "output_tokens": 528,
                "cached_tokens": 2145,
                "wall_ms": 7973
              },
              "judge_usage": {
                "input_tokens": 15267,
                "output_tokens": 972,
                "cached_tokens": 12425,
                "wall_ms": 18566
              }
            },
            {
              "draw_index": 1,
              "generation_hash": "c2248b92091f2d510922c39791ea89516068f04782b71c69531a54ddfda88f90",
              "status": "measured",
              "samples": [
                0.6,
                0.6,
                0.6
              ],
              "judge_sample_hashes": [
                "53f3ee0e6d867daa8b24cacfde49c6f56012ec66cb60bc1e421ba6709e83e729",
                "9ada8db0cccf1f3a139cf00a6114163418d4b0282ad3534c4f353075556a6d64",
                "264abb3fcb41d8b47bebccb251812deb96af5f65470c014995d8e7151cceade5"
              ],
              "mean": 0.6,
              "stddev": 0,
              "stop_reason": "end_turn",
              "truncated": false,
              "reported_model": "claude-opus-5-5",
              "usage": {
                "input_tokens": 3724,
                "output_tokens": 721,
                "cached_tokens": 3722,
                "wall_ms": 9540
              },
              "judge_usage": {
                "input_tokens": 15534,
                "output_tokens": 1276,
                "cached_tokens": 12603,
                "wall_ms": 22279
              }
            },
            {
              "draw_index": 2,
              "generation_hash": "294a63179dd624e32b3121cb9b74d9eb7d738ee60ddbccc783713c2cb8dad2db",
              "status": "measured",
              "samples": [
                0.58,
                0.55,
                0.6
              ],
              "judge_sample_hashes": [
                "9046a910c3a44c9ca9183d0d502504433596472ffdae13566e9a61d9bb72f553",
                "7fe7bc76ee9fde13943fee009cf3c01617335109fbf3d1d31265d3ed2c9b84c3",
                "cc506ae22d6f5f3635b7576e1f0284f914d1b394fd656194693c37512a4fe182"
              ],
              "mean": 0.576667,
              "stddev": 0.025166,
              "stop_reason": "end_turn",
              "truncated": false,
              "reported_model": "claude-opus-5-5",
              "usage": {
                "input_tokens": 3724,
                "output_tokens": 587,
                "cached_tokens": 3722,
                "wall_ms": 8364
              },
              "judge_usage": {
                "input_tokens": 15366,
                "output_tokens": 1238,
                "cached_tokens": 12491,
                "wall_ms": 20881
              }
            }
          ]
        }
      }
    ],
    "aggregates": {
      "with_skill": {
        "case_count": 1,
        "pass_count": 0,
        "borderline_count": 1,
        "mean_score": 0.657778,
        "stddev": null
      },
      "baseline": {
        "case_count": 1,
        "pass_count": 0,
        "borderline_count": 0,
        "mean_score": 0.586667,
        "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": 0.657778,
    "baseline_score": 0.586667,
    "delta": 0.071111,
    "delta_uncertainty": null,
    "delta_uncertainty_unavailable": "single_case"
  },
  "verification_level": "TESTED",
  "receipt_hash": "51a3d724dd17be440b7cee2d53d753165ea3c362ee8d1aa07b35630125e10b7b",
  "economics": {
    "basis": "metered-equivalent",
    "surface": "claude-cli",
    "with_skill": {
      "call_count": 6,
      "mean_input_tokens": 7187,
      "mean_output_tokens": 857.17,
      "mean_cost_usd_per_call": 0.045891,
      "median_wall_ms": 10856.5,
      "wall_ms_p25": 10773,
      "wall_ms_p75": 10882,
      "wall_ms_iqr": 109
    },
    "baseline": {
      "call_count": 3,
      "mean_input_tokens": 3724,
      "mean_output_tokens": 612,
      "mean_cost_usd_per_call": 0.027136,
      "median_wall_ms": 8364,
      "wall_ms_p25": null,
      "wall_ms_p75": null,
      "wall_ms_iqr": null
    },
    "skill_incremental_cost_usd_per_call": 0.018755,
    "skill_incremental_cost_usd_per_1k_calls": 18.755,
    "output_tokens_delta": 245.17,
    "median_wall_ms_delta": 2492.5,
    "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": 0.173316,
      "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"
    }
  }
}