{
  "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",
    "model_release_date": "2026-07-24",
    "provider": "anthropic",
    "surface": "claude-cli",
    "runner_version": "0.11.2",
    "date_utc": "2026-09-24T07:40:41.370Z",
    "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",
      "reported_models": [
        "claude-opus-5"
      ],
      "isolation": "same-user"
    },
    "pricing_snapshot": {
      "frozen_at": "2026-09-24T07:40:41.370Z",
      "source": "config/models.json",
      "currency": "USD",
      "models": {
        "claude-opus-5": {
          "input_per_mtok": 5,
          "output_per_mtok": 25,
          "registered": true
        },
        "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": "pass",
        "score": 0.801428,
        "mean": 0.801428,
        "stddev": 0.099142,
        "samples": [
          0.87,
          0.85,
          0.87
        ],
        "generation_hash": "4ae269d33831032647e00faeb2d308028768f4181167c24ed6d659842c1dcfe2",
        "judge_sample_hashes": [
          "8c6b3c4a624890cd163e7d782a0135c0063efa4adf7783973a4bcc0c4f07fee1",
          "4cf210ef48726721507a11b32babb368dbc9d878e9c9686899e71722a1f56ce9",
          "8988711fd1e061365383bb876f638016e8c506f0aac204becc378beaff2612f0"
        ],
        "threshold": 0.7,
        "reason": "Meets (a)-(d): removes restating comments, commented-out legacy line and bare TODO; added comments state mutation contract, non-idempotence and stale-total gotcha rather than restating code.",
        "judge": {
          "model_id": "claude-opus-5-5",
          "rubric_hash": "b38f01e6671fc0b28a94de550fe70f879134a3168b35e5b1fb53a9a6773478b0"
        },
        "judge_usage": {
          "input_tokens": 16836,
          "output_tokens": 835,
          "cached_tokens": 13471,
          "wall_ms": 18751
        },
        "generation": {
          "n_planned": 3,
          "n_drawn": 7,
          "n_measured": 7,
          "n_unmeasured": 0,
          "stopping_reason": "stabilised",
          "mean": 0.801428,
          "sd": 0.099142,
          "judge_sd_mean": 0.012905,
          "variance_ratio": 7.682449,
          "variance_ratio_unavailable": null,
          "n_truncated": 0,
          "draws": [
            {
              "draw_index": 0,
              "generation_hash": "ef528cf258194e94ce99de38322a2e1f5668bdc778f45c05376ef287e2697c24",
              "status": "measured",
              "samples": [
                0.78,
                0.8,
                0.8
              ],
              "judge_sample_hashes": [
                "dbf574eec07a3c3f232ec49673369cb612e0c4826218ee5d5ffc9e44bb7cfd24",
                "cb2020898730df8acd1603101ead4cce87062da5f579e56bfa5c2916c60c87bb",
                "e5f26af2471d2dbbb2511006caa4b3c2b953963a1795ea67789fe211c1c2f9d8"
              ],
              "mean": 0.793333,
              "stddev": 0.011547,
              "stop_reason": "end_turn",
              "truncated": false,
              "reported_model": "claude-opus-5",
              "usage": {
                "input_tokens": 8187,
                "output_tokens": 1778,
                "cached_tokens": 540,
                "wall_ms": 26147
              },
              "judge_usage": {
                "input_tokens": 16491,
                "output_tokens": 1015,
                "cached_tokens": 13241,
                "wall_ms": 18262
              }
            },
            {
              "draw_index": 1,
              "generation_hash": "742f7ecc47fd7cf15794d117ba1353a5393ce5bbfb84697ae244b60888f0e942",
              "status": "measured",
              "samples": [
                0.6,
                0.55,
                0.6
              ],
              "judge_sample_hashes": [
                "f207b3078660b443bdd9ade4c9834fba8043052064f64eba9ae7ae724e2198e4",
                "5dc7f205204fb850bdc5ad2ab9fb51ebb01e55191b5f44931ce257621c114844",
                "08884e6b5a6671c2ea6447aec8f5918388334afba2398d1a9fc784f7b1c16f16"
              ],
              "mean": 0.583333,
              "stddev": 0.028868,
              "stop_reason": "end_turn",
              "truncated": false,
              "reported_model": "claude-opus-5",
              "usage": {
                "input_tokens": 8187,
                "output_tokens": 1520,
                "cached_tokens": 8185,
                "wall_ms": 23218
              },
              "judge_usage": {
                "input_tokens": 16692,
                "output_tokens": 1239,
                "cached_tokens": 13375,
                "wall_ms": 19196
              }
            },
            {
              "draw_index": 2,
              "generation_hash": "661199bc80a3dceed98f089253756a54be0ad74a9876f7f0f1af8582772bfadb",
              "status": "measured",
              "samples": [
                0.84,
                0.85,
                0.84
              ],
              "judge_sample_hashes": [
                "21f9b04ebc9e7293fa602f467e4f52ef4be18278270acf79a3fc8af4980a4123",
                "22ecf831807a0fde18066379bc081e722f5cb3b13313ce8a5fa42da681a5c485",
                "684f1421beab67c03f2ab23e6662d597bfcc792c85ca09c01d645395f9215faa"
              ],
              "mean": 0.843333,
              "stddev": 0.005774,
              "stop_reason": "end_turn",
              "truncated": false,
              "reported_model": "claude-opus-5",
              "usage": {
                "input_tokens": 8187,
                "output_tokens": 1672,
                "cached_tokens": 8185,
                "wall_ms": 23322
              },
              "judge_usage": {
                "input_tokens": 16563,
                "output_tokens": 1001,
                "cached_tokens": 13289,
                "wall_ms": 18215
              }
            },
            {
              "draw_index": 3,
              "generation_hash": "cde921d8ac7d8a282ae1ade3ed0cec146e723d4dca343502b0dbfe4b421a858a",
              "status": "measured",
              "samples": [
                0.86,
                0.87,
                0.84
              ],
              "judge_sample_hashes": [
                "9b195b989a68dcdfa34c30bb1c5fdabf3d1b0647f351b8162b7965efe9857934",
                "f83597bc6a17ed95921abbae06d7b8e3362a5b07f33aa9891d51785a02103a42",
                "6a46bf03fb05be384aea96b85344d61a54467af65f7c233f698d37ddb809885b"
              ],
              "mean": 0.856667,
              "stddev": 0.015275,
              "stop_reason": "end_turn",
              "truncated": false,
              "reported_model": "claude-opus-5",
              "usage": {
                "input_tokens": 8187,
                "output_tokens": 2038,
                "cached_tokens": 8185,
                "wall_ms": 27330
              },
              "judge_usage": {
                "input_tokens": 16866,
                "output_tokens": 821,
                "cached_tokens": 13491,
                "wall_ms": 17047
              }
            },
            {
              "draw_index": 4,
              "generation_hash": "1a4c5aea8933db28b884ad26aa485460a6241b660f21c92577a4343e01cd1bb5",
              "status": "measured",
              "samples": [
                0.85,
                0.85,
                0.85
              ],
              "judge_sample_hashes": [
                "de0e9728a785c20d67ce1a7a52c2a0530b52c331fc42add1efcded52ac58cc45",
                "2169c95c54af4c2d7811c3d5cc3b865fec774f9f830f74abbdb43f7104d93b2a",
                "7ff1fc44e267e24bbeec08d8e78166145e1d616daac14c7519ecc3cc2863c5f5"
              ],
              "mean": 0.85,
              "stddev": 0,
              "stop_reason": "end_turn",
              "truncated": false,
              "reported_model": "claude-opus-5",
              "usage": {
                "input_tokens": 8187,
                "output_tokens": 1789,
                "cached_tokens": 8185,
                "wall_ms": 24360
              },
              "judge_usage": {
                "input_tokens": 15930,
                "output_tokens": 920,
                "cached_tokens": 12867,
                "wall_ms": 18269
              }
            },
            {
              "draw_index": 5,
              "generation_hash": "d61412a3b394e49f6c8ad13a06609006d69fc526fa909a7503758ad5ce4ba187",
              "status": "measured",
              "samples": [
                0.83,
                0.83,
                0.8
              ],
              "judge_sample_hashes": [
                "5246bebfe413167f411384b6a1a59a069c804212a2d72acd2fa55507a61c56a5",
                "73bcadc7a0395a9ab725410c9f1898da4e2f286a77244f7caa734eca91b774a4",
                "9aaf345b1f3cb4dbb1cc79ab22c72d8cf93afb76d2eade1d6516f18bcfd78be6"
              ],
              "mean": 0.82,
              "stddev": 0.017321,
              "stop_reason": "end_turn",
              "truncated": false,
              "reported_model": "claude-opus-5",
              "usage": {
                "input_tokens": 8187,
                "output_tokens": 1469,
                "cached_tokens": 8185,
                "wall_ms": 24377
              },
              "judge_usage": {
                "input_tokens": 16182,
                "output_tokens": 1005,
                "cached_tokens": 13035,
                "wall_ms": 18796
              }
            },
            {
              "draw_index": 6,
              "generation_hash": "4ae269d33831032647e00faeb2d308028768f4181167c24ed6d659842c1dcfe2",
              "status": "measured",
              "samples": [
                0.87,
                0.85,
                0.87
              ],
              "judge_sample_hashes": [
                "8c6b3c4a624890cd163e7d782a0135c0063efa4adf7783973a4bcc0c4f07fee1",
                "4cf210ef48726721507a11b32babb368dbc9d878e9c9686899e71722a1f56ce9",
                "8988711fd1e061365383bb876f638016e8c506f0aac204becc378beaff2612f0"
              ],
              "mean": 0.863333,
              "stddev": 0.011547,
              "stop_reason": "end_turn",
              "truncated": false,
              "reported_model": "claude-opus-5",
              "usage": {
                "input_tokens": 8187,
                "output_tokens": 1876,
                "cached_tokens": 8185,
                "wall_ms": 27536
              },
              "judge_usage": {
                "input_tokens": 16836,
                "output_tokens": 835,
                "cached_tokens": 13471,
                "wall_ms": 18751
              }
            }
          ]
        }
      },
      {
        "id": "comment-intent-not-implementation",
        "mode": "baseline",
        "outcome": "fail",
        "score": 0.516667,
        "mean": 0.516667,
        "stddev": 0.109291,
        "samples": [
          0.55,
          0.5,
          0.55
        ],
        "generation_hash": "27a7abe8938979818de5f1152730d36a9717f186fdea8f1eec45b1a08e0de2c3",
        "judge_sample_hashes": [
          "1bc9f55b45eae751decbb7633c96835e2e8bfab302d509cea57a3cac2314b9c6",
          "aab36945c1ea8e812c46bdc773558cec9b2c10950343268b85b97e61732fa789",
          "4309897f764218382a5f1a85d52eeee6ef7dfee73a3fd30e8220498b87de34e2"
        ],
        "threshold": 0.7,
        "reason": "Meets (a) and (b), but misses (c) because the TODO is kept rather than handled or removed; (d) is weakened because 'flat 10% off' restates 0.9.",
        "judge": {
          "model_id": "claude-opus-5-5",
          "rubric_hash": "b38f01e6671fc0b28a94de550fe70f879134a3168b35e5b1fb53a9a6773478b0"
        },
        "judge_usage": {
          "input_tokens": 16011,
          "output_tokens": 1272,
          "cached_tokens": 12921,
          "wall_ms": 21201
        },
        "generation": {
          "n_planned": 3,
          "n_drawn": 3,
          "n_measured": 3,
          "n_unmeasured": 0,
          "stopping_reason": "min_reached",
          "mean": 0.516667,
          "sd": 0.109291,
          "judge_sd_mean": 0.019245,
          "variance_ratio": 5.67893,
          "variance_ratio_unavailable": null,
          "n_truncated": 0,
          "draws": [
            {
              "draw_index": 0,
              "generation_hash": "88b0158403bbe83999e24b529a20a13a6405df8810a44723a5568a539f07dbdb",
              "status": "measured",
              "samples": [
                0.6,
                0.6,
                0.65
              ],
              "judge_sample_hashes": [
                "571756ddd94aa65ef2f90e16a585475a01818c3f0d3a17511dee778acf4098e4",
                "fb726404d9f660357235156f841976326ba50ded85d1c759c916bb183e14eabc",
                "6aab98ab2209406af321eca2ab4cd2f73e4cb6675f9101a7813445bd773880e0"
              ],
              "mean": 0.616667,
              "stddev": 0.028868,
              "stop_reason": "end_turn",
              "truncated": false,
              "reported_model": "claude-opus-5",
              "usage": {
                "input_tokens": 4724,
                "output_tokens": 1366,
                "cached_tokens": 3150,
                "wall_ms": 20894
              },
              "judge_usage": {
                "input_tokens": 16002,
                "output_tokens": 1237,
                "cached_tokens": 12915,
                "wall_ms": 21674
              }
            },
            {
              "draw_index": 1,
              "generation_hash": "ad51564ae49a988ef98c442d911a3556c861c1f0fecee8ec9867cc8306f70362",
              "status": "measured",
              "samples": [
                0.4,
                0.4,
                0.4
              ],
              "judge_sample_hashes": [
                "2231c5dccd67ad9f365a83cd938b1588b007fb281555e9df40203cf70b2d68c0",
                "e153995ca9b01393b73016c00dedb6d44bcd9b46328cc95a6fed95317259baf0",
                "a7c4f5696c49732688ee4d05d0707331ded045bae3d5be70fc1558022a0e79a5"
              ],
              "mean": 0.4,
              "stddev": 0,
              "stop_reason": "end_turn",
              "truncated": false,
              "reported_model": "claude-opus-5",
              "usage": {
                "input_tokens": 4724,
                "output_tokens": 999,
                "cached_tokens": 4722,
                "wall_ms": 15021
              },
              "judge_usage": {
                "input_tokens": 15798,
                "output_tokens": 1121,
                "cached_tokens": 12779,
                "wall_ms": 18858
              }
            },
            {
              "draw_index": 2,
              "generation_hash": "27a7abe8938979818de5f1152730d36a9717f186fdea8f1eec45b1a08e0de2c3",
              "status": "measured",
              "samples": [
                0.55,
                0.5,
                0.55
              ],
              "judge_sample_hashes": [
                "1bc9f55b45eae751decbb7633c96835e2e8bfab302d509cea57a3cac2314b9c6",
                "aab36945c1ea8e812c46bdc773558cec9b2c10950343268b85b97e61732fa789",
                "4309897f764218382a5f1a85d52eeee6ef7dfee73a3fd30e8220498b87de34e2"
              ],
              "mean": 0.533333,
              "stddev": 0.028868,
              "stop_reason": "end_turn",
              "truncated": false,
              "reported_model": "claude-opus-5",
              "usage": {
                "input_tokens": 4724,
                "output_tokens": 1076,
                "cached_tokens": 4722,
                "wall_ms": 16114
              },
              "judge_usage": {
                "input_tokens": 16011,
                "output_tokens": 1272,
                "cached_tokens": 12921,
                "wall_ms": 21201
              }
            }
          ]
        }
      }
    ],
    "aggregates": {
      "with_skill": {
        "case_count": 1,
        "pass_count": 1,
        "borderline_count": 0,
        "mean_score": 0.801428,
        "stddev": null
      },
      "baseline": {
        "case_count": 1,
        "pass_count": 0,
        "borderline_count": 0,
        "mean_score": 0.516667,
        "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": 0.801428,
    "baseline_score": 0.516667,
    "delta": 0.284761,
    "delta_uncertainty": null,
    "delta_uncertainty_unavailable": "single_case"
  },
  "verification_level": "TESTED",
  "receipt_hash": "dd464b6678e1c87c0840ce7261a5a35b8be373f5bca27a69f475e55671355457",
  "economics": {
    "basis": "metered-equivalent",
    "surface": "claude-cli",
    "with_skill": {
      "call_count": 7,
      "mean_input_tokens": 8187,
      "mean_output_tokens": 1734.57,
      "mean_cost_usd_per_call": 0.084299,
      "median_wall_ms": 24377,
      "wall_ms_p25": 23322,
      "wall_ms_p75": 27330,
      "wall_ms_iqr": 4008
    },
    "baseline": {
      "call_count": 3,
      "mean_input_tokens": 4724,
      "mean_output_tokens": 1147,
      "mean_cost_usd_per_call": 0.052295,
      "median_wall_ms": 16114,
      "wall_ms_p25": null,
      "wall_ms_p75": null,
      "wall_ms_iqr": null
    },
    "skill_incremental_cost_usd_per_call": 0.032004,
    "skill_incremental_cost_usd_per_1k_calls": 32.004,
    "output_tokens_delta": 587.57,
    "median_wall_ms_delta": 8263,
    "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": 0.173528,
      "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"
    }
  }
}