GHSA-XFQ4-8HF9-5H43
Vulnerability from github – Published: 2026-11-19 16:05 – Updated: 2026-11-19 16:05What's wrong
During restore, transparency_logged is recomputed as 1 if (toml_logged is not False and bundle_has_rekor) else 0, where bundle_has_rekor comes only from a rekor.uuid field inside signature_bundle. That derivation is wrong in both directions.
First, the rekor block is metadata attached after signing (attach_rekor_entry in mareforma/signing/rekor.py leaves the payload and signatures untouched), so it is not covered by the Ed25519 signature. A hand-edited claims.toml can add "rekor": {"uuid": "x"} to any claim's bundle and forge the "publicly witnessed by Rekor" state. The claim signature still verifies, and restore never cross-checks the block against [rekor_inclusions] or the log. Tampering with claims.toml is the exact threat model restore defends against; _verify_claim_signatures_on_restore checks only payload and signatures.
Second, a signed claim in a graph that never used Rekor is written with transparency_logged=1 (the column default when Rekor is disabled) but restores as 0 because its bundle has no rekor block. The convergence gate requires transparency_logged = 1, so post-restore peers can no longer converge to REPLICATED. With no rekor_url configured, refresh_unsigned() is a no-op and the flag stays 0 forever. Observed on the honest path: signed non-Rekor claim restores from transparency_logged=1 to 0, and a subsequent peer stays at PRELIMINARY instead of converging.
Severity and impact
High. An attacker who edits claims.toml can mark claims Rekor-witnessed and REPLICATED-eligible without them ever having been logged. Meanwhile the common honest configuration (signed, no Rekor) has its trust state silently degraded by every catastrophic-loss restore, with no recovery path.
Location
mareforma/db/restore.py:376readstoml_loggedfrom the TOML and probes the bundle for arekor.uuidmareforma/db/restore.py:385computesresolved_transparencyfrom the unsigned rekor block alonemareforma/db/core.py:1401the REPLICATED convergence gate requiresc.transparency_logged = 1
Fix direction
Do not derive transparency_logged from an unsigned envelope field. When a rekor block is present, require a matching, cryptographically verified [rekor_inclusions] entry (via verify_rekor_inclusion against the pinned log public key) whose hashedrekord spec.data.hash equals the SHA-256 of the claim's signed payload before setting 1; otherwise set 0.
When the claim is signed and neither the bundle nor the graph carries Rekor coordinates, restore the write-time value: the TOML omits transparency_logged unless it was 0, so treat an absent flag plus no rekor bundle as 1 (the column default for non-Rekor writes).
Do red/green TDD: add a regression test that fails on main, then make it pass.
Quality bar
SOLID, DRY, KISS. Clean, readable, well-tested code. No new errors, warnings, or test failures; the full suite stays green on all supported Python versions.
{
"affected": [
{
"package": {
"ecosystem": "PyPI",
"name": "mareforma"
},
"ranges": [
{
"events": [
{
"introduced": "0.3.0"
},
{
"fixed": "0.3.8"
}
],
"type": "ECOSYSTEM"
}
]
}
],
"aliases": [
"CVE-2026-71534"
],
"database_specific": {
"cwe_ids": [
"CWE-345"
],
"github_reviewed": true,
"github_reviewed_at": "2026-11-19T16:05:07Z",
"nvd_published_at": "2026-11-18T23:17:21Z",
"severity": "MODERATE"
},
"details": "## What\u0027s wrong\n\nDuring restore, `transparency_logged` is recomputed as `1 if (toml_logged is not False and bundle_has_rekor) else 0`, where `bundle_has_rekor` comes only from a `rekor.uuid` field inside `signature_bundle`. That derivation is wrong in both directions.\n\nFirst, the rekor block is metadata attached after signing (`attach_rekor_entry` in `mareforma/signing/rekor.py` leaves the payload and signatures untouched), so it is not covered by the Ed25519 signature. A hand-edited `claims.toml` can add `\"rekor\": {\"uuid\": \"x\"}` to any claim\u0027s bundle and forge the \"publicly witnessed by Rekor\" state. The claim signature still verifies, and restore never cross-checks the block against `[rekor_inclusions]` or the log. Tampering with `claims.toml` is the exact threat model restore defends against; `_verify_claim_signatures_on_restore` checks only payload and signatures.\n\nSecond, a signed claim in a graph that never used Rekor is written with `transparency_logged=1` (the column default when Rekor is disabled) but restores as 0 because its bundle has no rekor block. The convergence gate requires `transparency_logged = 1`, so post-restore peers can no longer converge to REPLICATED. With no `rekor_url` configured, `refresh_unsigned()` is a no-op and the flag stays 0 forever. Observed on the honest path: signed non-Rekor claim restores from `transparency_logged=1` to 0, and a subsequent peer stays at PRELIMINARY instead of converging.\n\n## Severity and impact\n\n**High.** An attacker who edits `claims.toml` can mark claims Rekor-witnessed and REPLICATED-eligible without them ever having been logged. Meanwhile the common honest configuration (signed, no Rekor) has its trust state silently degraded by every catastrophic-loss restore, with no recovery path.\n\n## Location\n\n- `mareforma/db/restore.py:376` reads `toml_logged` from the TOML and probes the bundle for a `rekor.uuid`\n- `mareforma/db/restore.py:385` computes `resolved_transparency` from the unsigned rekor block alone\n- `mareforma/db/core.py:1401` the REPLICATED convergence gate requires `c.transparency_logged = 1`\n\n## Fix direction\n\nDo not derive `transparency_logged` from an unsigned envelope field. When a rekor block is present, require a matching, cryptographically verified `[rekor_inclusions]` entry (via `verify_rekor_inclusion` against the pinned log public key) whose hashedrekord `spec.data.hash` equals the SHA-256 of the claim\u0027s signed payload before setting 1; otherwise set 0.\n\nWhen the claim is signed and neither the bundle nor the graph carries Rekor coordinates, restore the write-time value: the TOML omits `transparency_logged` unless it was 0, so treat an absent flag plus no rekor bundle as 1 (the column default for non-Rekor writes).\n\nDo red/green TDD: add a regression test that fails on `main`, then make it pass.\n\n## Quality bar\n\nSOLID, DRY, KISS. Clean, readable, well-tested code. No new errors, warnings, or test failures; the full suite stays green on all supported Python versions.",
"id": "GHSA-xfq4-8hf9-5h43",
"modified": "2026-11-19T16:05:07Z",
"published": "2026-11-19T16:05:07Z",
"references": [
{
"type": "WEB",
"url": "https://github.com/mareforma/mareforma/security/advisories/GHSA-xfq4-8hf9-5h43"
},
{
"type": "ADVISORY",
"url": "https://nvd.nist.gov/vuln/detail/CVE-2026-71534"
},
{
"type": "WEB",
"url": "https://github.com/mareforma/mareforma/pull/58"
},
{
"type": "WEB",
"url": "https://github.com/mareforma/mareforma/commit/cccbf4eafd3c34b20b1686f12c0cad76c87566ae"
},
{
"type": "PACKAGE",
"url": "https://github.com/mareforma/mareforma"
},
{
"type": "WEB",
"url": "https://github.com/mareforma/mareforma/releases/tag/v0.3.8"
}
],
"schema_version": "1.4.0",
"severity": [
{
"score": "CVSS:4.0/AV:L/AC:L/AT:N/PR:L/UI:N/VC:N/VI:H/VA:L/SC:N/SI:N/SA:N",
"type": "CVSS_V4"
}
],
"summary": "Mareforma: transparency_logged trusts the unsigned rekor block"
}
Sightings
| Author | Source | Type | Date | Other |
|---|
Nomenclature
- Seen: The vulnerability was mentioned, discussed, or observed by the user.
- Confirmed: The vulnerability has been validated from an analyst's perspective.
- Published Proof of Concept: A public proof of concept is available for this vulnerability.
- Exploited: The vulnerability was observed as exploited by the user who reported the sighting.
- Patched: The vulnerability was observed as successfully patched by the user who reported the sighting.
- Not exploited: The vulnerability was not observed as exploited by the user who reported the sighting.
- Not confirmed: The user expressed doubt about the validity of the vulnerability.
- Not patched: The vulnerability was not observed as successfully patched by the user who reported the sighting.
The approach is described in our paper Mapping CVEs to MITRE ATT&CK Techniques: A Curated Gold-Set Classifier and the Limits of LLM-Assisted Label Expansion.
Browse all ATT&CK techniques and the vulnerabilities related to each.
Related by attack behaviour
Vulnerabilities whose description is nearest to this one in the vector space of the CIRCL/vulnerability-attack-technique-biencoder model. This is a similarity search over the bi-encoder space (plain cosine), not a classification, and it has no measured accuracy.