proofbundle 4.0.0__tar.gz → 5.0.0__tar.gz

{proofbundle-4.0.0/src/proofbundle.egg-info → proofbundle-5.0.0}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: proofbundle
-Version: 4.0.0
+Version: 5.0.0
 Summary: Emit and verify portable cryptographic evidence bundles, offline: Ed25519 + RFC 6962 Merkle + optional SD-JWT.
 Author: Konrad Gruszka
 License: MIT
@@ -51,7 +51,7 @@ Requires-Dist: jsonschema>=4; extra == "test"
 Requires-Dist: sd-jwt>=0.10; extra == "test"
 Requires-Dist: rfc8785>=0.1.4; extra == "test"
 Provides-Extra: inspect
-Requires-Dist: inspect_ai<0.4,>=0.3.112; extra == "inspect"
+Requires-Dist: inspect_ai<=0.3.260,>=0.3.112; extra == "inspect"
 Provides-Extra: dev
 Requires-Dist: pytest>=7; extra == "dev"
 Requires-Dist: ruff<0.17,>=0.5; extra == "dev"
@@ -62,7 +62,7 @@ Requires-Dist: hypothesis>=6; extra == "dev"
 Requires-Dist: rfc8785>=0.1.4; extra == "dev"
 Requires-Dist: sd-jwt>=0.10; extra == "dev"
 Requires-Dist: PyYAML>=6; extra == "dev"
-Requires-Dist: inspect_ai<0.4,>=0.3.112; extra == "dev"
+Requires-Dist: inspect_ai<=0.3.260,>=0.3.112; extra == "dev"
 Dynamic: license-file
 
 <div align="center">
@@ -74,12 +74,6 @@ Dynamic: license-file
 
 <h1>proofbundle</h1>
 
-**AI eval results need receipts.**
-
-Turn an AI evaluation result into one portable, offline-verifiable receipt. It proves *who signed
-these exact bytes* and *that nothing changed since* — not that the number is true. Ed25519 + RFC 6962
-Merkle, one file, no server, no network.
-
 [![CI](https://github.com/b7n0de/proofbundle/actions/workflows/ci.yml/badge.svg)](https://github.com/b7n0de/proofbundle/actions/workflows/ci.yml)
 [![demo reproducible](https://github.com/b7n0de/proofbundle/actions/workflows/demo-reproducible.yml/badge.svg)](https://github.com/b7n0de/proofbundle/actions/workflows/demo-reproducible.yml)
 [![PyPI](https://img.shields.io/pypi/v/proofbundle.svg)](https://pypi.org/project/proofbundle/)
@@ -90,10 +84,48 @@ Merkle, one file, no server, no network.
 
 Scorecard **6.5/10** — [what the four zeros mean, in one sentence each](#what-the-scorecard-badge-says-including-the-parts-that-are-low)
 
+**AI eval results need receipts.**
+
+Turn an AI evaluation result into one portable, offline-verifiable receipt. It proves *who signed
+these exact bytes* and *that nothing changed since* — not that the number is true. Ed25519 + RFC 6962
+Merkle, one file, no server, no network.
+
 **Reviewing this for adoption?** Start with the 30-minute adversarial audit path: **[docs/REVIEWERS.md](https://github.com/b7n0de/proofbundle/blob/main/docs/REVIEWERS.md)**.
 
 </div>
 
+## Install
+
+```bash
+pip install proofbundle                 # core: offline verify + plain emit (two deps: cryptography, rfc8785)
+pip install "proofbundle[eval]"          # + eval receipts, prereg, and the demo (RFC 8785 JCS canonicalizer)
+pip install "proofbundle[inspect]"      # inspect_ai adapter + hook
+pip install "proofbundle[pq]"           # verify ML-DSA-44 (post-quantum) witness cosignatures
+```
+
+Requires Python 3.10+. The verify path never rolls its own crypto — Ed25519 comes from
+`cryptography`; Merkle hashing is RFC 6962.
+
+## 60-second try (offline)
+
+```bash
+pip install "proofbundle[eval]"
+proofbundle demo   # honest receipt => OK, six tampers each => FAILED, sample swap caught
+```
+
+The demo runs entirely in memory and exits non-zero if any tamper slips through, so it doubles as a
+self-test.
+
+```bash
+# verify a real hosted receipt without writing any code — the verify runs fully offline:
+curl -fsSL https://raw.githubusercontent.com/b7n0de/proofbundle/main/examples/example_bundle.json -o receipt.json
+proofbundle verify receipt.json        # CRYPTO: OK   (exit 0 ok · 1 fail · 2 malformed · 3 policy)
+```
+
+Emit your own receipt, apply a trust policy, start from a shipped template, or run the Inspect-native
+path (METR Task Standard / UK-AISI ecosystem, mockllm, no API key): **[docs/DEMO.md](https://github.com/b7n0de/proofbundle/blob/main/docs/DEMO.md)** ·
+Inspect walkthrough **[docs/INSPECT_HAPPY_PATH.md](https://github.com/b7n0de/proofbundle/blob/main/docs/INSPECT_HAPPY_PATH.md)**.
+
 ## What the Scorecard badge says, including the parts that are low
 
 The badge is live, so it will move. Measured 2026-08-07, re-measured 2026-08-10 with identical
@@ -128,26 +160,6 @@ measured and not yet addressed.
 Publishing a middling number with its causes is the point. A project that sells evidence cannot
 withhold its own.
 
-## 60-second try (offline)
-
-```bash
-pip install "proofbundle[eval]"
-proofbundle demo   # honest receipt => OK, six tampers each => FAILED, sample swap caught
-```
-
-The demo runs entirely in memory and exits non-zero if any tamper slips through, so it doubles as a
-self-test.
-
-```bash
-# verify a real hosted receipt without writing any code — the verify runs fully offline:
-curl -fsSL https://raw.githubusercontent.com/b7n0de/proofbundle/main/examples/example_bundle.json -o receipt.json
-proofbundle verify receipt.json        # CRYPTO: OK   (exit 0 ok · 1 fail · 2 malformed · 3 policy)
-```
-
-Emit your own receipt, apply a trust policy, start from a shipped template, or run the Inspect-native
-path (METR Task Standard / UK-AISI ecosystem, mockllm, no API key): **[docs/DEMO.md](https://github.com/b7n0de/proofbundle/blob/main/docs/DEMO.md)** ·
-Inspect walkthrough **[docs/INSPECT_HAPPY_PATH.md](https://github.com/b7n0de/proofbundle/blob/main/docs/INSPECT_HAPPY_PATH.md)**.
-
 ## The problem
 
 Every AI eval number you read — a safety benchmark, a capability score, a leaderboard entry — is an
@@ -236,18 +248,6 @@ version history (see also [CHANGELOG.md](https://github.com/b7n0de/proofbundle/b
   separation (executor ≠ decision maker) — never a claim that the decision was correct.
   [decision-receipt.md](https://github.com/b7n0de/proofbundle/blob/main/docs/predicates/decision-receipt.md) · [action-outcome.md](https://github.com/b7n0de/proofbundle/blob/main/docs/predicates/action-outcome.md)
 
-## Install
-
-```bash
-pip install proofbundle                 # core: offline verify + plain emit (two deps: cryptography, rfc8785)
-pip install "proofbundle[eval]"          # + eval receipts, prereg, and the demo (RFC 8785 JCS canonicalizer)
-pip install "proofbundle[inspect]"      # inspect_ai adapter + hook
-pip install "proofbundle[pq]"           # verify ML-DSA-44 (post-quantum) witness cosignatures
-```
-
-Requires Python 3.10+. The verify path never rolls its own crypto — Ed25519 comes from
-`cryptography`; Merkle hashing is RFC 6962.
-
 ## Post-quantum posture (honest)
 
 proofbundle is **not** "quantum-safe" as a whole. Its hash-based layers (SHA-256, RFC 6962 / 9162
@@ -264,7 +264,8 @@ a post-quantum *payload* signature is on the roadmap. Detail: [docs/ANCHORS.md](
 If proofbundle helped your evaluation pipeline, please cite it. Machine-readable metadata is in
 [`CITATION.cff`](https://github.com/b7n0de/proofbundle/blob/main/CITATION.cff). The archival software record is on Zenodo under concept
 DOI [10.5281/zenodo.21110642](https://doi.org/10.5281/zenodo.21110642); the Technical Note (design write-up) under concept DOI
-[10.5281/zenodo.21230466](https://doi.org/10.5281/zenodo.21230466), also linked from [b7n0de.com/proofbundle](https://b7n0de.com/proofbundle).
+[10.5281/zenodo.21230466](https://doi.org/10.5281/zenodo.21230466) — version 4.0.0 of the Note is
+[10.5281/zenodo.22004295](https://doi.org/10.5281/zenodo.22004295) — also linked from [b7n0de.com/proofbundle](https://b7n0de.com/proofbundle).
 
 ## Docs
 

{proofbundle-4.0.0 → proofbundle-5.0.0}/README.md

@@ -7,12 +7,6 @@
 
 <h1>proofbundle</h1>
 
-**AI eval results need receipts.**
-
-Turn an AI evaluation result into one portable, offline-verifiable receipt. It proves *who signed
-these exact bytes* and *that nothing changed since* — not that the number is true. Ed25519 + RFC 6962
-Merkle, one file, no server, no network.
-
 [![CI](https://github.com/b7n0de/proofbundle/actions/workflows/ci.yml/badge.svg)](https://github.com/b7n0de/proofbundle/actions/workflows/ci.yml)
 [![demo reproducible](https://github.com/b7n0de/proofbundle/actions/workflows/demo-reproducible.yml/badge.svg)](https://github.com/b7n0de/proofbundle/actions/workflows/demo-reproducible.yml)
 [![PyPI](https://img.shields.io/pypi/v/proofbundle.svg)](https://pypi.org/project/proofbundle/)
@@ -23,10 +17,48 @@ Merkle, one file, no server, no network.
 
 Scorecard **6.5/10** — [what the four zeros mean, in one sentence each](#what-the-scorecard-badge-says-including-the-parts-that-are-low)
 
+**AI eval results need receipts.**
+
+Turn an AI evaluation result into one portable, offline-verifiable receipt. It proves *who signed
+these exact bytes* and *that nothing changed since* — not that the number is true. Ed25519 + RFC 6962
+Merkle, one file, no server, no network.
+
 **Reviewing this for adoption?** Start with the 30-minute adversarial audit path: **[docs/REVIEWERS.md](https://github.com/b7n0de/proofbundle/blob/main/docs/REVIEWERS.md)**.
 
 </div>
 
+## Install
+
+```bash
+pip install proofbundle                 # core: offline verify + plain emit (two deps: cryptography, rfc8785)
+pip install "proofbundle[eval]"          # + eval receipts, prereg, and the demo (RFC 8785 JCS canonicalizer)
+pip install "proofbundle[inspect]"      # inspect_ai adapter + hook
+pip install "proofbundle[pq]"           # verify ML-DSA-44 (post-quantum) witness cosignatures
+```
+
+Requires Python 3.10+. The verify path never rolls its own crypto — Ed25519 comes from
+`cryptography`; Merkle hashing is RFC 6962.
+
+## 60-second try (offline)
+
+```bash
+pip install "proofbundle[eval]"
+proofbundle demo   # honest receipt => OK, six tampers each => FAILED, sample swap caught
+```
+
+The demo runs entirely in memory and exits non-zero if any tamper slips through, so it doubles as a
+self-test.
+
+```bash
+# verify a real hosted receipt without writing any code — the verify runs fully offline:
+curl -fsSL https://raw.githubusercontent.com/b7n0de/proofbundle/main/examples/example_bundle.json -o receipt.json
+proofbundle verify receipt.json        # CRYPTO: OK   (exit 0 ok · 1 fail · 2 malformed · 3 policy)
+```
+
+Emit your own receipt, apply a trust policy, start from a shipped template, or run the Inspect-native
+path (METR Task Standard / UK-AISI ecosystem, mockllm, no API key): **[docs/DEMO.md](https://github.com/b7n0de/proofbundle/blob/main/docs/DEMO.md)** ·
+Inspect walkthrough **[docs/INSPECT_HAPPY_PATH.md](https://github.com/b7n0de/proofbundle/blob/main/docs/INSPECT_HAPPY_PATH.md)**.
+
 ## What the Scorecard badge says, including the parts that are low
 
 The badge is live, so it will move. Measured 2026-08-07, re-measured 2026-08-10 with identical
@@ -61,26 +93,6 @@ measured and not yet addressed.
 Publishing a middling number with its causes is the point. A project that sells evidence cannot
 withhold its own.
 
-## 60-second try (offline)
-
-```bash
-pip install "proofbundle[eval]"
-proofbundle demo   # honest receipt => OK, six tampers each => FAILED, sample swap caught
-```
-
-The demo runs entirely in memory and exits non-zero if any tamper slips through, so it doubles as a
-self-test.
-
-```bash
-# verify a real hosted receipt without writing any code — the verify runs fully offline:
-curl -fsSL https://raw.githubusercontent.com/b7n0de/proofbundle/main/examples/example_bundle.json -o receipt.json
-proofbundle verify receipt.json        # CRYPTO: OK   (exit 0 ok · 1 fail · 2 malformed · 3 policy)
-```
-
-Emit your own receipt, apply a trust policy, start from a shipped template, or run the Inspect-native
-path (METR Task Standard / UK-AISI ecosystem, mockllm, no API key): **[docs/DEMO.md](https://github.com/b7n0de/proofbundle/blob/main/docs/DEMO.md)** ·
-Inspect walkthrough **[docs/INSPECT_HAPPY_PATH.md](https://github.com/b7n0de/proofbundle/blob/main/docs/INSPECT_HAPPY_PATH.md)**.
-
 ## The problem
 
 Every AI eval number you read — a safety benchmark, a capability score, a leaderboard entry — is an
@@ -169,18 +181,6 @@ version history (see also [CHANGELOG.md](https://github.com/b7n0de/proofbundle/b
   separation (executor ≠ decision maker) — never a claim that the decision was correct.
   [decision-receipt.md](https://github.com/b7n0de/proofbundle/blob/main/docs/predicates/decision-receipt.md) · [action-outcome.md](https://github.com/b7n0de/proofbundle/blob/main/docs/predicates/action-outcome.md)
 
-## Install
-
-```bash
-pip install proofbundle                 # core: offline verify + plain emit (two deps: cryptography, rfc8785)
-pip install "proofbundle[eval]"          # + eval receipts, prereg, and the demo (RFC 8785 JCS canonicalizer)
-pip install "proofbundle[inspect]"      # inspect_ai adapter + hook
-pip install "proofbundle[pq]"           # verify ML-DSA-44 (post-quantum) witness cosignatures
-```
-
-Requires Python 3.10+. The verify path never rolls its own crypto — Ed25519 comes from
-`cryptography`; Merkle hashing is RFC 6962.
-
 ## Post-quantum posture (honest)
 
 proofbundle is **not** "quantum-safe" as a whole. Its hash-based layers (SHA-256, RFC 6962 / 9162
@@ -197,7 +197,8 @@ a post-quantum *payload* signature is on the roadmap. Detail: [docs/ANCHORS.md](
 If proofbundle helped your evaluation pipeline, please cite it. Machine-readable metadata is in
 [`CITATION.cff`](https://github.com/b7n0de/proofbundle/blob/main/CITATION.cff). The archival software record is on Zenodo under concept
 DOI [10.5281/zenodo.21110642](https://doi.org/10.5281/zenodo.21110642); the Technical Note (design write-up) under concept DOI
-[10.5281/zenodo.21230466](https://doi.org/10.5281/zenodo.21230466), also linked from [b7n0de.com/proofbundle](https://b7n0de.com/proofbundle).
+[10.5281/zenodo.21230466](https://doi.org/10.5281/zenodo.21230466) — version 4.0.0 of the Note is
+[10.5281/zenodo.22004295](https://doi.org/10.5281/zenodo.22004295) — also linked from [b7n0de.com/proofbundle](https://b7n0de.com/proofbundle).
 
 ## Docs
 

{proofbundle-4.0.0 → proofbundle-5.0.0}/conformance/manifest.json

@@ -58,6 +58,16 @@
     "relation/statement-retracts-declared-unresolved",
     "relation/statement-supersedes-verified",
     "relation/statement-malformed",
-    "relation/wrong-payloadtype-target"
+    "relation/wrong-payloadtype-target",
+    "provenance/version-status-reported",
+    "provenance/version-status-not-reported",
+    "provenance/version-status-not-bound",
+    "provenance/version-status-reported-without-field",
+    "provenance/version-status-not-reported-with-field",
+    "provenance/version-status-missing-reason",
+    "provenance/version-status-unknown-literal",
+    "provenance/version-status-unrelated-status-field",
+    "provenance/version-status-absent-legacy-receipt",
+    "provenance/version-status-lookalike-version-field"
   ]
 }

proofbundle-5.0.0/conformance/provenance/version-status-absent-legacy-receipt/case.json

@@ -0,0 +1,15 @@
+{
+  "caseId": "provenance-version-status-absent-legacy-receipt",
+  "kind": "provenance_version_status",
+  "input": "provenance.json",
+  "attribution": "proofbundle 5.0.0 — reported-version status (jury lens 1, 2026-08-26)",
+  "expected": {
+    "issues": []
+  },
+  "specRefs": [
+    "EVAL_CLAIM.md",
+    "CONFORMANCE.md",
+    "CHANGELOG.md"
+  ],
+  "rationale": "A receipt issued before 5.0.0 carries NO status at all and stays valid. The change is additive by requirement, not by accident: demanding a status would invalidate every receipt ever issued. This vector locks that in, so a later reading cannot quietly turn absence into a rejection."
+}

proofbundle-5.0.0/conformance/provenance/version-status-absent-legacy-receipt/provenance.json

@@ -0,0 +1,4 @@
+{
+  "harness": "lm-eval",
+  "harness_version": "0.4.12"
+}

proofbundle-5.0.0/conformance/provenance/version-status-lookalike-version-field/case.json

@@ -0,0 +1,15 @@
+{
+  "caseId": "provenance-version-status-lookalike-version-field",
+  "kind": "provenance_version_status",
+  "input": "provenance.json",
+  "attribution": "proofbundle 5.0.0 — reported-version status (jury lens 1 round 2, 2026-08-26)",
+  "expected": {
+    "issues": []
+  },
+  "specRefs": [
+    "EVAL_CLAIM.md",
+    "CONFORMANCE.md",
+    "CHANGELOG.md"
+  ],
+  "rationale": "`schema_version` ends in `_version` but is NOT a harness-reported field. Its status — even a malformed one — must produce NO finding, while the real harness field beside it is still checked. The rule governs a NAMED set, not every name that looks like a version: for a verifier a false finding (a valid receipt made to look non-conformant) is the worse error direction, and a suffix rule over-matches by construction."
+}

proofbundle-5.0.0/conformance/provenance/version-status-lookalike-version-field/provenance.json

@@ -0,0 +1,6 @@
+{
+  "schema_version_status": "v1",
+  "schema_version": "0.1",
+  "harness_version": "0.4.12",
+  "harness_version_status": "reported"
+}

proofbundle-5.0.0/conformance/provenance/version-status-missing-reason/case.json

@@ -0,0 +1,17 @@
+{
+  "caseId": "provenance-version-status-missing-reason",
+  "kind": "provenance_version_status",
+  "input": "provenance.json",
+  "attribution": "proofbundle 5.0.0 — reported-version status (deep-gate iteration 5, finding F-C)",
+  "expected": {
+    "issues": [
+      "harness_version_status=not_reported without harness_version_status_reason (reason is mandatory)"
+    ]
+  },
+  "specRefs": [
+    "EVAL_CLAIM.md",
+    "CONFORMANCE.md",
+    "CHANGELOG.md"
+  ],
+  "rationale": "A non-`reported` status without its reason moves the ambiguity instead of closing it — which is the whole point of the rule. The reason is mandatory and its absence is non-conformant."
+}

proofbundle-5.0.0/conformance/provenance/version-status-missing-reason/provenance.json

@@ -0,0 +1,3 @@
+{
+  "harness_version_status": "not_reported"
+}

proofbundle-5.0.0/conformance/provenance/version-status-not-bound/case.json

@@ -0,0 +1,15 @@
+{
+  "caseId": "provenance-version-status-not-bound",
+  "kind": "provenance_version_status",
+  "input": "provenance.json",
+  "attribution": "proofbundle 5.0.0 — reported-version status (deep-gate iteration 5, finding F-C)",
+  "expected": {
+    "issues": []
+  },
+  "specRefs": [
+    "EVAL_CLAIM.md",
+    "CONFORMANCE.md",
+    "CHANGELOG.md"
+  ],
+  "rationale": "No harness version is bound for this format at all. This is the OTHER meaning that bare absence used to carry, and it is now a distinct, named state."
+}

proofbundle-5.0.0/conformance/provenance/version-status-not-bound/provenance.json

@@ -0,0 +1,5 @@
+{
+  "harness": "custom-runner",
+  "harness_version_status": "not_bound",
+  "harness_version_status_reason": "this format carries no harness version; the adapter binds none"
+}

proofbundle-5.0.0/conformance/provenance/version-status-not-reported/case.json

@@ -0,0 +1,15 @@
+{
+  "caseId": "provenance-version-status-not-reported",
+  "kind": "provenance_version_status",
+  "input": "provenance.json",
+  "attribution": "proofbundle 5.0.0 — reported-version status (deep-gate iteration 5, finding F-C)",
+  "expected": {
+    "issues": []
+  },
+  "specRefs": [
+    "EVAL_CLAIM.md",
+    "CONFORMANCE.md",
+    "CHANGELOG.md"
+  ],
+  "rationale": "The harness RAN and reported no version. The version field stays ABSENT — writing a value the harness never reported is forbidden — and the status plus its mandatory reason say so explicitly. Before 5.0.0 this receipt was byte-indistinguishable from one where nobody bound the field at all."
+}

proofbundle-5.0.0/conformance/provenance/version-status-not-reported/provenance.json

@@ -0,0 +1,5 @@
+{
+  "harness": "lm-eval",
+  "harness_version_status": "not_reported",
+  "harness_version_status_reason": "lm-eval did not write `lm_eval_version` into results.json"
+}

proofbundle-5.0.0/conformance/provenance/version-status-not-reported-with-field/case.json

@@ -0,0 +1,17 @@
+{
+  "caseId": "provenance-version-status-not-reported-with-field",
+  "kind": "provenance_version_status",
+  "input": "provenance.json",
+  "attribution": "proofbundle 5.0.0 — reported-version status (deep-gate iteration 5, finding F-C)",
+  "expected": {
+    "issues": [
+      "harness_version_status=not_reported but harness_version is present ('0.4.12')"
+    ]
+  },
+  "specRefs": [
+    "EVAL_CLAIM.md",
+    "CONFORMANCE.md",
+    "CHANGELOG.md"
+  ],
+  "rationale": "The mirror contradiction: the status denies a report while the value sits right beside it. Both directions must be caught, otherwise the pair can lie in one of them."
+}

proofbundle-5.0.0/conformance/provenance/version-status-not-reported-with-field/provenance.json

@@ -0,0 +1,5 @@
+{
+  "harness_version": "0.4.12",
+  "harness_version_status": "not_reported",
+  "harness_version_status_reason": "x"
+}

proofbundle-5.0.0/conformance/provenance/version-status-reported/case.json

@@ -0,0 +1,15 @@
+{
+  "caseId": "provenance-version-status-reported",
+  "kind": "provenance_version_status",
+  "input": "provenance.json",
+  "attribution": "proofbundle 5.0.0 — reported-version status (deep-gate iteration 5, finding F-C)",
+  "expected": {
+    "issues": []
+  },
+  "specRefs": [
+    "EVAL_CLAIM.md",
+    "CONFORMANCE.md",
+    "CHANGELOG.md"
+  ],
+  "rationale": "The harness reported a version: the field is present and the status says `reported`. No reason is required — a reported version explains itself. This is the baseline case; an implementation that flags it has made the status itself a finding."
+}

proofbundle-5.0.0/conformance/provenance/version-status-reported/provenance.json

@@ -0,0 +1,5 @@
+{
+  "harness": "lm-eval",
+  "harness_version": "0.4.12",
+  "harness_version_status": "reported"
+}

proofbundle-5.0.0/conformance/provenance/version-status-reported-without-field/case.json

@@ -0,0 +1,17 @@
+{
+  "caseId": "provenance-version-status-reported-without-field",
+  "kind": "provenance_version_status",
+  "input": "provenance.json",
+  "attribution": "proofbundle 5.0.0 — reported-version status (deep-gate iteration 5, finding F-C)",
+  "expected": {
+    "issues": [
+      "harness_version_status=reported but harness_version is absent"
+    ]
+  },
+  "specRefs": [
+    "EVAL_CLAIM.md",
+    "CONFORMANCE.md",
+    "CHANGELOG.md"
+  ],
+  "rationale": "A status that claims `reported` while the field it speaks about is absent MUST be rejected. Without this case a verifier could be told anything by the status alone."
+}

proofbundle-5.0.0/conformance/provenance/version-status-reported-without-field/provenance.json

@@ -0,0 +1,3 @@
+{
+  "harness_version_status": "reported"
+}

proofbundle-5.0.0/conformance/provenance/version-status-unknown-literal/case.json

@@ -0,0 +1,17 @@
+{
+  "caseId": "provenance-version-status-unknown-literal",
+  "kind": "provenance_version_status",
+  "input": "provenance.json",
+  "attribution": "proofbundle 5.0.0 — reported-version status (deep-gate iteration 5, finding F-C)",
+  "expected": {
+    "issues": [
+      "harness_version_status='maybe' is not one of ['reported', 'not_reported', 'not_bound']"
+    ]
+  },
+  "specRefs": [
+    "EVAL_CLAIM.md",
+    "CONFORMANCE.md",
+    "CHANGELOG.md"
+  ],
+  "rationale": "Only the three English literals are conformant. An open vocabulary would let an implementation invent a fourth state and call a receipt conformant."
+}

proofbundle-5.0.0/conformance/provenance/version-status-unknown-literal/provenance.json

@@ -0,0 +1,3 @@
+{
+  "harness_version_status": "maybe"
+}

proofbundle-5.0.0/conformance/provenance/version-status-unrelated-status-field/case.json

@@ -0,0 +1,15 @@
+{
+  "caseId": "provenance-version-status-unrelated-status-field",
+  "kind": "provenance_version_status",
+  "input": "provenance.json",
+  "attribution": "proofbundle 5.0.0 — reported-version status (jury lens 1, 2026-08-26)",
+  "expected": {
+    "issues": []
+  },
+  "specRefs": [
+    "EVAL_CLAIM.md",
+    "CONFORMANCE.md",
+    "CHANGELOG.md"
+  ],
+  "rationale": "A provenance block may carry OTHER fields ending in `_status` (run_status, scorer_status). They are not harness-reported version fields and MUST NOT produce a finding. A verifier that invents a finding on a field it was never asked about makes a valid receipt look non-conformant — the failure this product exists against, pointed the other way. The rule follows the field suffix `_version`, not a hard-coded list, so a future `foo_version` is covered and nothing else is ever touched."
+}

proofbundle-5.0.0/conformance/provenance/version-status-unrelated-status-field/provenance.json

@@ -0,0 +1,6 @@
+{
+  "run_status": "active",
+  "scorer_status": "llm",
+  "harness_version": "0.4.12",
+  "harness_version_status": "reported"
+}

{proofbundle-4.0.0 → proofbundle-5.0.0}/conformance/run_conformance.py

@@ -293,9 +293,50 @@ def _check_relation_statement(case: dict, case_dir: pathlib.Path, *, require_anc
     return _check_relation(case, case_dir, verb="relation-statement")
 
 
+def _check_provenance_version_status(case: dict, case_dir: pathlib.Path, *,
+                                     require_anchors: bool = False) -> dict:
+    """A provenance block is checked against the reported-version status rule (v5.0.0).
+
+    WHY THIS KIND EXISTS. A harness-reported version field used to be simply ABSENT when the
+    harness reported none, so absence carried two meanings at once — "the harness reported
+    nothing" and "nobody bound this field". For a verifier that ambiguity is the failure class
+    the product exists against. v5.0.0 adds an explicit three-value status beside each such
+    field; the corpus is the original authority for that rule, so the rule needs cases here and
+    not only unit tests.
+
+    The case declares `expected.issues`: the exact list of problem sentences the verifier must
+    produce for that provenance block. An EMPTY list means the block is consistent. Declaring
+    the exact list (rather than only a count) is deliberate — a case that merely asserted "some
+    problem" would pass on the wrong problem.
+    """
+    cid = case.get("caseId", str(case_dir))
+    exp = case.get("expected") or {}
+    if "issues" not in exp:
+        return _fail(cid, "provenance_version_status case under-declares its expectations "
+                          "(fail-closed): missing `issues`")
+    name = case.get("input") or "provenance.json"
+    if pathlib.Path(name).is_absolute() or ".." in pathlib.PurePosixPath(name).parts:
+        return _fail(cid, f"input {name!r} escapes the case directory")
+    try:
+        prov = json.loads((case_dir / name).read_text())
+    except Exception as e:                                   # noqa: BLE001
+        return _fail(cid, f"{name} unreadable ({type(e).__name__}: {e})")
+    if not isinstance(prov, dict):
+        return _fail(cid, f"{name} must contain a JSON object")
+    sys.path.insert(0, str(ROOT.parent / "src"))
+    from proofbundle.adapters._provenance import version_status_issues  # noqa: PLC0415
+    got = sorted(version_status_issues(prov))
+    want = sorted(str(x) for x in exp["issues"])
+    if got != want:
+        return _fail(cid, f"issues {got!r} != expected {want!r}")
+    return {"caseId": cid, "ok": True,
+            "detail": f"{len(got)} issue(s) as expected"}
+
+
 _DISPATCH = {"decision_crossimpl": _check_decision_crossimpl, "native_bundle": _check_native_bundle,
              "decision_relation": _check_decision_relation, "outcome_relation": _check_outcome_relation,
-             "relation_statement": _check_relation_statement}
+             "relation_statement": _check_relation_statement,
+             "provenance_version_status": _check_provenance_version_status}
 
 
 def run(*, require_anchors: bool = False) -> int: