{
  "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": "2026-09-24T07:37:30.838Z",
    "registry": "registered",
    "transcripts": "hashes-only",
    "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-opus-5",
      "reported_models": [
        "claude-opus-5"
      ],
      "isolation": "same-user"
    },
    "pricing_snapshot": {
      "frozen_at": "2026-09-24T07:37:30.838Z",
      "source": "config/models.json",
      "currency": "USD",
      "models": {
        "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": 0.806667,
        "mean": 0.806667,
        "stddev": 0.109358,
        "samples": [
          0.87,
          0.87,
          0.87
        ],
        "generation_hash": "4ae269d33831032647e00faeb2d308028768f4181167c24ed6d659842c1dcfe2",
        "judge_sample_hashes": [
          "cefe42496dffe91950d4e70610820ea08b983820fee3436111d9656196d0d665",
          "5a449383b8a092b32cff94b17df45e5f7cd343bc7aaaca0fb155d10e7c31df57",
          "c68e285faff00dba1470c0e84fb7367ca92a9f211eaa96ee0268e236c1e056ca"
        ],
        "threshold": 0.7,
        "reason": "Removes all three restating comments and the commented-out legacy line, drops the bare TODO with justification, and adds genuinely useful contract/idempotence/staleness comments that explain intent rather than restate code.",
        "judge": {
          "model_id": "claude-opus-5",
          "rubric_hash": "b38f01e6671fc0b28a94de550fe70f879134a3168b35e5b1fb53a9a6773478b0"
        },
        "judge_usage": {
          "input_tokens": 19836,
          "output_tokens": 1674,
          "cached_tokens": 16476,
          "wall_ms": 27496
        },
        "generation": {
          "n_planned": 3,
          "n_drawn": 7,
          "n_measured": 7,
          "n_unmeasured": 0,
          "stopping_reason": "stabilised",
          "mean": 0.806667,
          "sd": 0.109358,
          "judge_sd_mean": 0.016034,
          "variance_ratio": 6.820382,
          "variance_ratio_unavailable": null,
          "n_truncated": 0,
          "draws": [
            {
              "draw_index": 0,
              "generation_hash": "ef528cf258194e94ce99de38322a2e1f5668bdc778f45c05376ef287e2697c24",
              "status": "measured",
              "samples": [
                0.86,
                0.85,
                0.85
              ],
              "judge_sample_hashes": [
                "556f631cf71086e0b3aa731e098f3a49bf871eee553eeaadfc6d24dfb2af3516",
                "4c8f1b338289843bad35443d8b38f49c590a952577d98450d05aac5075e1c4b7",
                "2967443f1b889bf12ca920a07d8009e27e5a974739c289ce72d7114baea4faca"
              ],
              "mean": 0.853333,
              "stddev": 0.005774,
              "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": 19491,
                "output_tokens": 1592,
                "cached_tokens": 16246,
                "wall_ms": 25860
              }
            },
            {
              "draw_index": 1,
              "generation_hash": "742f7ecc47fd7cf15794d117ba1353a5393ce5bbfb84697ae244b60888f0e942",
              "status": "measured",
              "samples": [
                0.55,
                0.6,
                0.55
              ],
              "judge_sample_hashes": [
                "868d6b35aefafd8e0cd428515ef961fc3873eb481c1c9d573658c3778af4a671",
                "8d560c4e59d89886172220e816aad3628955c1af1a80cf76ee898a8f2b91d21d",
                "9a81c9a3c88629f9df061f29bcac88d6410e7c8e3f66e7a0172aa65b9aa55a49"
              ],
              "mean": 0.566667,
              "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": 19692,
                "output_tokens": 2288,
                "cached_tokens": 16380,
                "wall_ms": 35425
              }
            },
            {
              "draw_index": 2,
              "generation_hash": "661199bc80a3dceed98f089253756a54be0ad74a9876f7f0f1af8582772bfadb",
              "status": "measured",
              "samples": [
                0.86,
                0.85,
                0.85
              ],
              "judge_sample_hashes": [
                "ebd0402cbb38b59d9176fade3c96521bbe1ccdb86ca0500660474548bd1b4719",
                "bb29127130c7aa28a340d901a6d168ccaff20d1e397b796c16cc468057b7650e",
                "89d0022f442daad233ea4a0c40985800886ceeb2f546a01825f3b30b221f3a48"
              ],
              "mean": 0.853333,
              "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": 19563,
                "output_tokens": 1413,
                "cached_tokens": 16294,
                "wall_ms": 24876
              }
            },
            {
              "draw_index": 3,
              "generation_hash": "cde921d8ac7d8a282ae1ade3ed0cec146e723d4dca343502b0dbfe4b421a858a",
              "status": "measured",
              "samples": [
                0.86,
                0.87,
                0.86
              ],
              "judge_sample_hashes": [
                "171802bb4bc9e50e0ade22469d0b7945aff2615b103483dca6266d1a7c45f9a0",
                "28b393f53c468f520a53fc3fb68a2e1598d2978700e33b5eecd12ca0b795c9c0",
                "6c6df3078621a38b179fe65b281f06da2f3c35b4689c1688338dc377b9fb43ca"
              ],
              "mean": 0.863333,
              "stddev": 0.005774,
              "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": 19866,
                "output_tokens": 1568,
                "cached_tokens": 16496,
                "wall_ms": 25889
              }
            },
            {
              "draw_index": 4,
              "generation_hash": "1a4c5aea8933db28b884ad26aa485460a6241b660f21c92577a4343e01cd1bb5",
              "status": "measured",
              "samples": [
                0.86,
                0.85,
                0.85
              ],
              "judge_sample_hashes": [
                "60581bc8656beaad3d8324da030681127db747d41b45f0b2e320af3a1e14f13d",
                "59f12256317014d89d37102437a7df08c1e60decb25f6f7e4bb352822a566348",
                "3d87935defcc0114a372c38781ae600ae9257aea4d670482b7f625257c079196"
              ],
              "mean": 0.853333,
              "stddev": 0.005774,
              "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": 18930,
                "output_tokens": 1736,
                "cached_tokens": 15872,
                "wall_ms": 33886
              }
            },
            {
              "draw_index": 5,
              "generation_hash": "d61412a3b394e49f6c8ad13a06609006d69fc526fa909a7503758ad5ce4ba187",
              "status": "measured",
              "samples": [
                0.78,
                0.73,
                0.85
              ],
              "judge_sample_hashes": [
                "364d30a1a3d4f35d6ef22d19c4b8c77f799af0bcb41f1fdf67492177b1706e86",
                "c7c3688597d63a83555162dabb5fd535f4175fb17dee5e255153f34b0719d35b",
                "e43fbc41a11dca05ea85b43913751fa68dc363311782607248d0df1de49ba3eb"
              ],
              "mean": 0.786667,
              "stddev": 0.060277,
              "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": 19182,
                "output_tokens": 2849,
                "cached_tokens": 16040,
                "wall_ms": 41613
              }
            },
            {
              "draw_index": 6,
              "generation_hash": "4ae269d33831032647e00faeb2d308028768f4181167c24ed6d659842c1dcfe2",
              "status": "measured",
              "samples": [
                0.87,
                0.87,
                0.87
              ],
              "judge_sample_hashes": [
                "cefe42496dffe91950d4e70610820ea08b983820fee3436111d9656196d0d665",
                "5a449383b8a092b32cff94b17df45e5f7cd343bc7aaaca0fb155d10e7c31df57",
                "c68e285faff00dba1470c0e84fb7367ca92a9f211eaa96ee0268e236c1e056ca"
              ],
              "mean": 0.87,
              "stddev": 0,
              "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": 19836,
                "output_tokens": 1674,
                "cached_tokens": 16476,
                "wall_ms": 27496
              }
            }
          ]
        }
      },
      {
        "id": "comment-intent-not-implementation",
        "mode": "baseline",
        "outcome": "fail",
        "score": 0.583333,
        "mean": 0.583333,
        "stddev": 0.045826,
        "samples": [
          0.6,
          0.58,
          0.6
        ],
        "generation_hash": "27a7abe8938979818de5f1152730d36a9717f186fdea8f1eec45b1a08e0de2c3",
        "judge_sample_hashes": [
          "9c9bff92e3fb8f34ea86122a82fe2a8124fe1ba67865e6c32431dd39c62ef3b3",
          "178b80d52f7150b56909977b4db644385c28463bcaa8e5564f94d183bae5e8ba",
          "4a65e74212a712afe9e38bc98a538237a6e06b77a2c49f68c8933b0f46e65fcf"
        ],
        "threshold": 0.7,
        "reason": "Meets (a), (b), and largely (d), but leaves the coupon TODO unresolved — merely reworded rather than implemented or removed — and `flat 10% off` restates the 0.9.",
        "judge": {
          "model_id": "claude-opus-5",
          "rubric_hash": "b38f01e6671fc0b28a94de550fe70f879134a3168b35e5b1fb53a9a6773478b0"
        },
        "judge_usage": {
          "input_tokens": 19011,
          "output_tokens": 2100,
          "cached_tokens": 15926,
          "wall_ms": 34221
        },
        "generation": {
          "n_planned": 3,
          "n_drawn": 3,
          "n_measured": 3,
          "n_unmeasured": 0,
          "stopping_reason": "min_reached",
          "mean": 0.583333,
          "sd": 0.045826,
          "judge_sd_mean": 0.02186,
          "variance_ratio": 2.09634,
          "variance_ratio_unavailable": null,
          "n_truncated": 0,
          "draws": [
            {
              "draw_index": 0,
              "generation_hash": "88b0158403bbe83999e24b529a20a13a6405df8810a44723a5568a539f07dbdb",
              "status": "measured",
              "samples": [
                0.6,
                0.62,
                0.65
              ],
              "judge_sample_hashes": [
                "d03084c23de1b429109317ea9630a1b475411adfdd699eae69d785a3bcd85f9e",
                "6893b1e8a7497631204c131a2c1d3f603a7ad5a41911c3de4895ffe5d1f0c14f",
                "b33671e2829926eb13a048af27014a636f8385ef2f67981e6b9408d9debe8eb7"
              ],
              "mean": 0.623333,
              "stddev": 0.025166,
              "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": 19002,
                "output_tokens": 1857,
                "cached_tokens": 15920,
                "wall_ms": 30307
              }
            },
            {
              "draw_index": 1,
              "generation_hash": "ad51564ae49a988ef98c442d911a3556c861c1f0fecee8ec9867cc8306f70362",
              "status": "measured",
              "samples": [
                0.55,
                0.55,
                0.5
              ],
              "judge_sample_hashes": [
                "d89f024202f62d82d5778cc1f587b6cfd9bc15ec8afd663d2e1076dfbdf0a30c",
                "54e024bcc7905005599d220c81df16891f5321ba2de5c0317cc0b0fb1e775fef",
                "1968256d9e7434b5fb481bb70159b53f31b018473f4b73ad7bee522e0584fc9b"
              ],
              "mean": 0.533333,
              "stddev": 0.028868,
              "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": 18798,
                "output_tokens": 2757,
                "cached_tokens": 15784,
                "wall_ms": 41081
              }
            },
            {
              "draw_index": 2,
              "generation_hash": "27a7abe8938979818de5f1152730d36a9717f186fdea8f1eec45b1a08e0de2c3",
              "status": "measured",
              "samples": [
                0.6,
                0.58,
                0.6
              ],
              "judge_sample_hashes": [
                "9c9bff92e3fb8f34ea86122a82fe2a8124fe1ba67865e6c32431dd39c62ef3b3",
                "178b80d52f7150b56909977b4db644385c28463bcaa8e5564f94d183bae5e8ba",
                "4a65e74212a712afe9e38bc98a538237a6e06b77a2c49f68c8933b0f46e65fcf"
              ],
              "mean": 0.593333,
              "stddev": 0.011547,
              "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": 19011,
                "output_tokens": 2100,
                "cached_tokens": 15926,
                "wall_ms": 34221
              }
            }
          ]
        }
      }
    ],
    "aggregates": {
      "with_skill": {
        "case_count": 1,
        "pass_count": 0,
        "borderline_count": 1,
        "mean_score": 0.806667,
        "stddev": null
      },
      "baseline": {
        "case_count": 1,
        "pass_count": 0,
        "borderline_count": 0,
        "mean_score": 0.583333,
        "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": 0.806667,
    "baseline_score": 0.583333,
    "delta": 0.223334,
    "delta_uncertainty": null,
    "delta_uncertainty_unavailable": "single_case"
  },
  "verification_level": "TESTED",
  "receipt_hash": "2d871adfaefa09c0cc5ab22a50dc7b32f34f264f1a1e22196085e0b6bde76575",
  "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": 0.288585,
      "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"
    }
  }
}