Capture harness, fixture verification, CI, and the public README

The site does no live inference. Rollouts are captured once against spark-1 and
replayed at their recorded wall-clock — a public demo with no auth cannot hold
an API key, and a recorded run can be scrubbed, permalinked, blind-compared and
verified in ways a live one cannot. What stops it being a video is that the
browser re-derives every number from the recorded moves.

verify_fixtures.py is the Python half of that: it replays every committed
fixture through the engine and reproduces its own rewards. All 16 land at
delta 0.0. A fixture that cannot be regenerated is a claim with no receipt.

First real measurement, thinking off, 8 seeds: solved 0/8. The model repeats
guesses it has already played, invents words (trape, slith, postt, boomy),
and contradicts its own feedback — consistency 0.09 to 0.17. That is the
published failure taxonomy showing up in our own data on the first run, and it
is why `consistency` is a reward component rather than a footnote.

A capture failure is recorded as a turn with a null reply, never dropped. A
capture that silently discarded failed turns would be reporting a better model
than the one that ran.

CI gates both halves and four things that fail silently in production: the word
lists must rebuild byte-identically, the prerendered routes must carry their own
baked og tags (crawlers do not run JS, so without them every shared link
previews as the homepage), no blob: URL may reach the bundle (the site's CSP has
no worker-src, so it falls back to default-src 'self' and a blob worker is
blocked with no error), and the conformance digest must match across languages.

Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_019mt6sHQHEnEYrJZvoMCJSB

public/traces/wordle/base-off-s4.json

@@ -0,0 +1,188 @@
+{
+  "runId": "base-off-s4",
+  "seed": 4,
+  "model": "brain-qwen38-dspark",
+  "capturedAt": "2026-08-28",
+  "rewards": {
+    "solved": 0.0,
+    "economy": 0.0,
+    "consistency": 0.09090909090909091
+  },
+  "metrics": {
+    "guesses_used": 5.0,
+    "rejected_replies": 6.0,
+    "reference_depth": 4.0,
+    "turns_granted": 6.0,
+    "inconsistent_guesses": 5.0
+  },
+  "truncated": false,
+  "outcome": "failed",
+  "answer": "divot",
+  "turns": [
+    {
+      "reply": "[crane]",
+      "reasoning": null,
+      "call": {
+        "promptTokens": 147,
+        "completionTokens": 5,
+        "reasoningTokens": null,
+        "durationMs": 952,
+        "finishReason": "stop"
+      },
+      "info": {
+        "guess": "crane",
+        "observation": "C R A N E\nX X X X X\n\nYou have 5 guesses left."
+      }
+    },
+    {
+      "reply": "[slith]",
+      "reasoning": null,
+      "call": {
+        "promptTokens": 184,
+        "completionTokens": 5,
+        "reasoningTokens": null,
+        "durationMs": 1501,
+        "finishReason": "stop"
+      },
+      "info": {
+        "guess": "slith",
+        "observation": "That guess was not accepted: 'slith' is not in the word list.\nIt did not use up a turn. You have 5 guesses left."
+      }
+    },
+    {
+      "reply": "[slant]",
+      "reasoning": null,
+      "call": {
+        "promptTokens": 235,
+        "completionTokens": 5,
+        "reasoningTokens": null,
+        "durationMs": 1363,
+        "finishReason": "stop"
+      },
+      "info": {
+        "guess": "slant",
+        "observation": "S L A N T\nX X X X G\n\nYou have 4 guesses left."
+      }
+    },
+    {
+      "reply": "[roast]",
+      "reasoning": null,
+      "call": {
+        "promptTokens": 272,
+        "completionTokens": 5,
+        "reasoningTokens": null,
+        "durationMs": 1609,
+        "finishReason": "stop"
+      },
+      "info": {
+        "guess": "roast",
+        "observation": "R O A S T\nX Y X X G\n\nYou have 3 guesses left."
+      }
+    },
+    {
+      "reply": "[toast]",
+      "reasoning": null,
+      "call": {
+        "promptTokens": 309,
+        "completionTokens": 4,
+        "reasoningTokens": null,
+        "durationMs": 1240,
+        "finishReason": "stop"
+      },
+      "info": {
+        "guess": "toast",
+        "observation": "T O A S T\nX Y X X G\n\nYou have 2 guesses left."
+      }
+    },
+    {
+      "reply": "[plout]",
+      "reasoning": null,
+      "call": {
+        "promptTokens": 345,
+        "completionTokens": 5,
+        "reasoningTokens": null,
+        "durationMs": 1613,
+        "finishReason": "stop"
+      },
+      "info": {
+        "guess": "plout",
+        "observation": "That guess was not accepted: 'plout' is not in the word list.\nIt did not use up a turn. You have 2 guesses left."
+      }
+    },
+    {
+      "reply": "[postt]",
+      "reasoning": null,
+      "call": {
+        "promptTokens": 396,
+        "completionTokens": 5,
+        "reasoningTokens": null,
+        "durationMs": 1349,
+        "finishReason": "stop"
+      },
+      "info": {
+        "guess": "postt",
+        "observation": "That guess was not accepted: 'postt' is not in the word list.\nIt did not use up a turn. You have 2 guesses left."
+      }
+    },
+    {
+      "reply": "[roost]",
+      "reasoning": null,
+      "call": {
+        "promptTokens": 447,
+        "completionTokens": 5,
+        "reasoningTokens": null,
+        "durationMs": 868,
+        "finishReason": "stop"
+      },
+      "info": {
+        "guess": "roost",
+        "observation": "R O O S T\nX Y X X G\n\nYou have 1 guess left."
+      }
+    },
+    {
+      "reply": "[plout]",
+      "reasoning": null,
+      "call": {
+        "promptTokens": 484,
+        "completionTokens": 5,
+        "reasoningTokens": null,
+        "durationMs": 946,
+        "finishReason": "stop"
+      },
+      "info": {
+        "guess": "plout",
+        "observation": "That guess was not accepted: 'plout' is not in the word list.\nIt did not use up a turn. You have 1 guess left."
+      }
+    },
+    {
+      "reply": "[toast]",
+      "reasoning": null,
+      "call": {
+        "promptTokens": 535,
+        "completionTokens": 4,
+        "reasoningTokens": null,
+        "durationMs": 498,
+        "finishReason": "stop"
+      },
+      "info": {
+        "guess": "toast",
+        "observation": "That guess was not accepted: 'toast' has already been guessed.\nIt did not use up a turn. You have 1 guess left."
+      }
+    },
+    {
+      "reply": "[most]",
+      "reasoning": null,
+      "call": {
+        "promptTokens": 582,
+        "completionTokens": 4,
+        "reasoningTokens": null,
+        "durationMs": 490,
+        "finishReason": "stop"
+      },
+      "info": {
+        "guess": "most",
+        "observation": "That guess was not accepted: 'most' is not 5 letters.\nIt did not use up a turn. You have 1 guess left."
+      }
+    }
+  ]
+}