Emissary CVE-2026-35583

The CVSS v3.1 score of 5.3 reflects a medium-severity vulnerability with network accessibility (AV:N) and low attack complexity (AC:L) from unauthenticated actors (PR:N, UI:N). The impact is limited to confidentiality (C:L) with no integrity or availability effects, consistent with a read-only information disclosure scenario. The EPSS exploitation probability is not provided in the available data, limiting dynamic risk quantification. The absence of CISA KEV status and publicly available proof-of-concept code suggests this is not an actively exploited vulnerability at the time of analysis. However, path traversal vulnerabilities affecting API endpoints are generally straightforward to exploit once understood, and the bypass techniques (URL encoding, double-encoding, Unicode normalization) are well-established attack methods. The real-world risk depends on the sensitivity of configuration files accessible through this endpoint; if configurations contain secrets, database credentials, or other sensitive data, the actual impact could exceed the CVSS score. Organizations running Emissary in network-exposed environments should treat this as a priority remediation item, though the low attack complexity and unauthenticated access vector make this a practical concern for any internet-facing deployment.

An unauthenticated attacker on the network discovers an exposed Emissary instance and attempts to read sensitive configuration files by crafting malicious requests to the /api/configuration/{name} endpoint. The attacker bypasses the blacklist validation by sending a request such as /api/configuration/%2e%2e%2fconfig.xml (URL-encoded parent directory traversal) or /api/configuration/%252e%252e%252fconfig.xml (double-encoded), which evades the simple character-based blacklist checks. The server processes the encoded request, normalizes it to the actual path traversal sequence, and returns configuration files from outside the intended directory, exposing database credentials, API keys, or other secrets stored in those files. No specialized tools or privileged access is required; any HTTP client can execute this attack.