Crypt::SecretBuffer CVE-2026-5086

Exploitation requires that the target application actually uses Crypt::SecretBuffer's comparison functionality to check attacker-influenced input against a stored secret (the description's example is storing and comparing plaintext passwords) - the vulnerable code path must be reachable and its timing observable by the attacker. Limiting factors: the attacker must be able to submit many comparison attempts and measure fine-grained timing, which network jitter degrades; the flaw yields information disclosure only (C:H, I:N, A:N) and does not by itself grant code execution; and if the deploying application does not use SecretBuffer for secret comparison, or wraps it behind rate limiting or hashing, exploitation is impractical. No authentication is required to attempt comparisons per the CVSS vector (PR:N), though the surrounding application may impose its own access controls.

The signals are mostly consistent toward LOW real-world urgency. The supplied CVSS 3.1 vector (AV:N/AC:L/PR:N/UI:N, C:H) scores 7.5 (High) driven entirely by confidentiality impact, but AC:L understates reality: practical timing attacks over a network require large numbers of samples and statistical filtering of jitter, so effective complexity is higher. EPSS is negligible (0.02%, 5th percentile), SSVC exploitation is 'none', it is not in CISA KEV, and no POC is referenced - all pointing away from imminent exploitation. SSVC does flag Automatable: yes and Technical Impact: partial, meaning that where the vulnerable comparison IS reachable, an attack could be scripted. The dominant limiting factor is exposure: exploitation depends on an application actually using SecretBuffer to compare attacker-influenced input against a secret, which is a usage pattern the description raises only as an example ('if... used to store and compare plaintext passwords'). Net: patch it as good hygiene, but it is not a drop-everything issue.

An attacker repeatedly submits candidate passwords to an application that uses Crypt::SecretBuffer (<0.019) to compare the submitted value against a stored plaintext secret, measuring the response time of each attempt. By observing that closer-matching guesses take measurably longer (or shorter) to reject, the attacker reconstructs the secret one position at a time. No public POC is referenced, and given network timing noise the attack requires many samples to be reliable.