MISP Guard CVE-2026-118173

Exploitation requires that MISP Guard be configured with a required-taxonomy rule and an allowed-tags list (for example TLP with tlp:clear allowed), and that the attacker can author or modify events or attributes in a feed source the guard proxies - including attributes nested within objects and shadow attributes, all of which the pre-1.4 check missed. The specific trigger is an event or attribute that simultaneously carries an allowed taxonomy tag and a non-allowed tag of the same required taxonomy; the pre-patch guard verified only that one matching allowed tag existed and never checked that every tag of that taxonomy was allowed. Limiting factors: the attacker needs privileged write access to feed content (PR:H), the guard must be deployed with required-taxonomy enforcement enabled, and the affected version must be below 1.4; there is no impact on the guard host's own confidentiality, integrity, or availability.

This is a moderate-severity data-handling bypass, not a high-urgency emergency. The vendor CVSS 4.0 score is 6.1 (Medium) with vector AV:N/AC:L/AT:N/PR:H/UI:N and the only impact set to SI:H (subsequent-system integrity), reflecting that the flaw breaks the integrity of the feed's data-classification policy rather than compromising the guard host itself. The decisive limiting signal is PR:H: exploitation requires an actor who can already author or modify MISP events in a feed source the guard proxies, so this is an abuse-of-trusted-contributor issue, not a remote-unauthenticated break. No EPSS score, KEV listing, or SSVC automatable/technical-impact values are present in the input, so widespread opportunistic exploitation cannot be inferred - the only provided priority signal is vuln.today's 'this_month' label, consistent with a real but non-emergency fix. The references contain a patch commit only; there is no public exploit identified at time of analysis. One cross-source note: the vendor models the impact as subsequent-system integrity (SI:H), but the practical effect described - higher-classification data (e.g., TLP:red/amber) reaching consumers cleared only for lower-classification data - is fundamentally a confidentiality/disclosure outcome for downstream feed consumers, which is why the tags label it 'Information Disclosure'; defenders should treat it as a classification-leak risk.

An organization runs MISP Guard with a required-taxonomy rule (e.g., TLP) that only permits tlp:clear to flow to a given feed consumer. A contributor with write access to the feed source authors an event carrying both tlp:clear (allowed) and tlp:red (not allowed); because the guard only confirmed that at least one taxonomy tag matched the allow-list, the event passes and the sensitive data is distributed to consumers who should never receive it. No public exploit code is identified; the fix is in the referenced vendor patch commit.