hippoley · Oct 7, 2026 · Oct 7, 2026 · Oct 7, 2026 · Oct 7, 2026 · Oct 7, 2026
diff --git a/README.md b/README.md
@@ -56,6 +56,38 @@ The same QEMU artifacts now feed three evidence layers: the Bluefin-specific con
 
 ---
 
+
+## Flagship boundary: green judge, contradicted product oracle
+
+A second flagship case shows the opposite failure mode: **green submitted evidence can still be wrong about product truth**.
+
+In [anthropics/claude-code#89404](https://github.com/anthropics/claude-code/pull/89404), the submitted validator suite reported 5/5 passing. Reviewer-supplied measurements showed two stronger facts:
+
+```text
+submitted judge       5/5 PASS
+multi-line regression suite stays green after extraction revert
+product oracle        claude plugin validate --json → REJECT
+oracle alignment      CONTRADICTED
+```
+
+CounterProof publishes this as an `ORACLE_CONTRADICTED_BY_PRODUCT` artifact. It is intentionally marked as **reviewer-supplied external-oracle evidence**; CounterProof does not claim it independently executed the Claude Code binary.
+
+This case complements Bluefin:
+
+```text
+Bluefin
+same oracle + controlled intervention + recovery
+→ positive causal witness
+
+Claude #89404
+green submitted judge + rejecting product oracle
+→ negative proof boundary
+```
+
+**[Read the oracle-disagreement artifact →](reality/claude-code-89404/ORACLE_DISAGREEMENT.md)** · **[Read the machine receipt →](examples/claim_matrix/receipts/claude-code-89404-oracle.json)**
+
+---
+
 ## Tested on real agent PRs
 
 CounterProof is not developed only against fixtures. New proof semantics are tested against public AI-assisted pull requests where a reviewer has a concrete reason not to trust a green check.

diff --git a/docs/REALITY_LAB.md b/docs/REALITY_LAB.md
@@ -14,7 +14,7 @@ The rule is simple:
 | [openai/codex-plugin-cc#731](https://github.com/openai/codex-plugin-cc/pull/731) | Do the changed tests prove the original fail-open bug, and do they also answer later reviewer concerns? | **WITNESSED + proof boundary** | The tests prove the original bug, but not later 0600/atomic-write concerns. Green evidence must not silently expand its claim. |
 | [openai/codex-plugin-cc#456](https://github.com/openai/codex-plugin-cc/pull/456) | What if the thing being fixed is the test harness/environment isolation itself? | **NOT WITNESSED + REVIEW REQUIRED** | Direct BASE/HEAD reproduction supports the bug, but ordinary witness is withheld because changed test support is part of the fix. |
 | [MetrolistGroup/Metrolist#4097](https://github.com/MetrolistGroup/Metrolist/pull/4097) | Does a BASE non-zero exit always count as regression evidence? | **INCONCLUSIVE** | No. The BASE failed during Kotlin test compilation, not at a behavioral assertion. This produced the structured `json-v1` result protocol. |
-| [anthropics/claude-code#89404](https://github.com/anthropics/claude-code/pull/89404) | What if the submitted tests are internally green but disagree with the product's authoritative parser? | **Open Reality Probe** | This is the emerging oracle-alignment problem: red→green evidence can still be weak if the judge does not represent product truth. |
+| [anthropics/claude-code#89404](https://github.com/anthropics/claude-code/pull/89404) | What if the submitted tests are internally green but disagree with the product's authoritative parser? | **[ORACLE CONTRADICTED](../reality/claude-code-89404/ORACLE_DISAGREEMENT.md)** | Reviewer-supplied product-oracle evidence shows `claude plugin validate --json` rejects fixtures the submitted validator accepts, while the submitted suite can remain 5/5 green even when the multi-line extraction behavior regresses. CounterProof publishes this as an external-oracle contradiction, not as an independently executed Claude runtime replay. |
 | [continuedev/continue#12576](https://github.com/continuedev/continue/pull/12576) | Can CounterProof replay a package-local Vitest regression inside a large JS/TS monorepo? | **[WITNESSED](../reality/continue-12576/WITNESS.md)** | The real PR exposed two product gaps first: `*.vitest.ts` discovery and package-local `core/node_modules` closure. After both fixes, HEAD was 5/5 PASS and BASE was 4/5 PASS, 1 FAIL; the BASE failure directly hits the legacy `tabAutocompleteModel` migration claim. |
 | [topoteretes/cognee#5161](https://github.com/topoteretes/cognee/pull/5161) | What if local safety tests pass but a human reviewer reproduces a broader end-to-end failure mode those tests never exercise? | **Human oracle CONTRADICTED** | The vector adapter refusal test is narrow; reviewer evidence shows graph deletion can happen before the refusal propagates. CounterProof records the human reproduction as an external oracle instead of pretending it ran PostgreSQL itself. |
 | [vercel/ai#17096](https://github.com/vercel/ai/pull/17096) | What if two newly added tests have different evidentiary value, while a stronger production claim lives outside the test suite? | **[WITNESSED + NOT WITNESSED + UNVERIFIED](../reality/vercel-ai-17096/CLAIM_MATRIX.md)** | Host→bridge sandbox forwarding is red→green; the new schema test passes on BASE and HEAD; real Claude Agent SDK enforcement remains an external production oracle. |

diff --git a/examples/claim_matrix/claude-code-89404.yml b/examples/claim_matrix/claude-code-89404.yml
@@ -7,8 +7,11 @@ claims:
     submitted_test_evidence: UNPROVEN
     oracle_alignment: CONTRADICTED
     oracle_probe: "claude plugin validate rejects fixtures that the submitted validator treats as valid"
-    oracle_source_url: "https://github.com/anthropics/claude-code/pull/89404"
-    note: "Reviewer evidence establishes product-oracle contradiction; CounterProof has not independently replayed BASE/HEAD for this row."
+    oracle_source_url: "https://github.com/anthropics/claude-code/pull/89404#pullrequestreview-5190939241"
+    receipt_file: "receipts/claude-code-89404-oracle.json"
+    receipt_expected_verdicts:
+      - ORACLE_CONTRADICTED_BY_PRODUCT
+    note: "Reviewer evidence establishes product-oracle contradiction; CounterProof has not independently executed the Claude binary for this row."
   - id: multiline-description-regression
     claim: "The submitted suite protects the multi-line description extraction behavior"
     submitted_test_evidence: UNPROVEN

diff --git a/examples/claim_matrix/receipts/claude-code-89404-oracle.json b/examples/claim_matrix/receipts/claude-code-89404-oracle.json
@@ -0,0 +1,34 @@
+{
+  "schema_version": 1,
+  "case": "anthropics/claude-code#89404",
+  "verdict": "ORACLE_CONTRADICTED_BY_PRODUCT",
+  "evidence_kind": "REVIEWER_SUPPLIED_EXTERNAL_ORACLE",
+  "source_pr": "https://github.com/anthropics/claude-code/pull/89404",
+  "candidate": {
+    "base_sha": "8b6ef81f636a7697e5ae2338428fa0b272993845",
+    "head_sha": "0989f29b0bca412edc58018c462cae7344bb1a2f"
+  },
+  "submitted_judge": {
+    "test": "plugins/plugin-dev/skills/agent-development/scripts/validate-agent.test.sh",
+    "reported_result": "5/5 PASS",
+    "limitation": "Reviewer reproduced that reverting the multi-line description extraction still leaves the suite green while the false <example> warning returns."
+  },
+  "authoritative_oracle": {
+    "command": "claude plugin validate --json",
+    "reported_versions": [
+      "2.1.268",
+      "2.1.270"
+    ],
+    "result": "REJECT",
+    "reason": "The product parser rejects fixtures that the submitted validator test treats as valid.",
+    "review_source": "https://github.com/anthropics/claude-code/pull/89404#pullrequestreview-5190939241",
+    "inline_source": "https://github.com/anthropics/claude-code/pull/89404#discussion_r3999827406",
+    "followup_source": "https://github.com/anthropics/claude-code/pull/89404#issuecomment-5555827628"
+  },
+  "conclusion": "The submitted judge is not aligned with the product parser for the claimed-valid fixtures. A green submitted suite must not be promoted to product-level proof.",
+  "does_not_claim": [
+    "CounterProof independently executed the Claude Code binary.",
+    "The entire pull request is invalid.",
+    "Every submitted regression assertion is false."
+  ]
+}
diff --git a/reality/claude-code-89404/ORACLE_DISAGREEMENT.md b/reality/claude-code-89404/ORACLE_DISAGREEMENT.md
@@ -0,0 +1,78 @@
+# Claude Code #89404 — oracle disagreement proof
+
+CounterProof uses this case as a flagship **negative proof boundary**.
+
+The submitted regression suite reports green, but reviewer-supplied evidence shows
+that the product's own parser rejects fixtures the submitted validator treats as
+valid.
+
+## Candidate identity
+
+- BASE: `8b6ef81f636a7697e5ae2338428fa0b272993845`
+- HEAD: `0989f29b0bca412edc58018c462cae7344bb1a2f`
+
+## Submitted judge
+
+`validate-agent.test.sh` reports 5/5 passing.
+
+A reviewer also reproduced that reverting the multi-line description extraction
+back to the original single-line form still leaves that suite green while the false
+`<example>` warning returns.
+
+That means the submitted judge does not actually guard the claimed multi-line
+description behavior.
+
+## Authoritative oracle
+
+Reviewer measurements used the product command:
+
+```text
+claude plugin validate --json
+```
+
+on Claude Code 2.1.268 and 2.1.270.
+
+The product parser rejected the agent fixtures that the submitted validator test
+treated as valid.
+
+Primary external evidence:
+
+- review: https://github.com/anthropics/claude-code/pull/89404#pullrequestreview-5190939241
+- inline parser finding: https://github.com/anthropics/claude-code/pull/89404#discussion_r3999827406
+- reviewer agreement on product-parser alignment: https://github.com/anthropics/claude-code/pull/89404#issuecomment-5555827628
+
+## CounterProof conclusion
+
+```text
+submitted judge       GREEN
+authoritative oracle  REJECT
+oracle alignment      CONTRADICTED
+product-level claim   CONTRADICTED
+```
+
+This artifact is intentionally labeled `REVIEWER_SUPPLIED_EXTERNAL_ORACLE`.
+
+CounterProof did **not** independently execute the Claude Code binary for this
+publication artifact. The proof is therefore about the documented oracle
+contradiction and its provenance, not a claim that CounterProof reproduced the
+product runtime itself.
+
+## Why this belongs beside Bluefin
+
+Bluefin demonstrates the positive case:
+
+```text
+same oracle
+CONTROL == REVERT != BAD
+→ causal witness
+```
+
+Claude #89404 demonstrates the equally important negative case:
+
+```text
+submitted judge green
+product oracle rejects
+→ do not promote green evidence into product truth
+```
+
+Together they define the product boundary more clearly than either example alone.
diff --git a/tests/unit/test_claim_matrix.py b/tests/unit/test_claim_matrix.py
@@ -598,3 +598,30 @@ def test_scancode_example_binds_db_backed_behavior_receipt():
     assert payload["claims"][0]["oracle_applicability"] == "APPLICABLE"
     assert payload["claims"][0]["oracle_alignment"] == "ALIGNED"
     assert payload["claims"][0]["overall_claim"] == "PROVEN"
+
+
+def test_claude_oracle_disagreement_binds_external_oracle_receipt():
+    from skill_factory.evolution.claim_matrix import load_claim_matrix
+
+    manifest = load_claim_matrix(Path("examples/claim_matrix/claude-code-89404.yml"))
+    payload = claim_matrix_to_dict(manifest)
+
+    first = payload["claims"][0]
+    assert first["overall_claim"] == "CONTRADICTED"
+    assert first["oracle_alignment"] == "CONTRADICTED"
+    assert first["receipt_observed_verdict"] == "ORACLE_CONTRADICTED_BY_PRODUCT"
+    assert first["receipt_case"] == "anthropics/claude-code#89404"
+
+    receipt_path = Path(
+        "examples/claim_matrix/receipts/claude-code-89404-oracle.json"
+    )
+    receipt = json.loads(receipt_path.read_text(encoding="utf-8"))
+    assert receipt["evidence_kind"] == "REVIEWER_SUPPLIED_EXTERNAL_ORACLE"
+    assert receipt["candidate"]["head_sha"] == (
+        "0989f29b0bca412edc58018c462cae7344bb1a2f"
+    )
+    assert receipt["authoritative_oracle"]["command"] == (
+        "claude plugin validate --json"
+    )
+    assert receipt["authoritative_oracle"]["result"] == "REJECT"
+    assert receipt["does_not_claim"]