GoClaw CVE-2026-15624

Requires a low-privilege authenticated session on the GoClaw application, as indicated by PR:L in the CVSS 4.0 vector - unauthenticated access alone is insufficient. The BytePlus video integration feature (bytePlusDownloadVideo) must be active and the invoke endpoint must be network-reachable by the authenticated attacker (AV:N). No special network positioning or race condition is required (AC:L, AT:N). Exploitation is constrained by the authentication requirement, which limits opportunistic mass exploitation. In deployments where the invoke endpoint is accessible only internally, external attackers must first obtain valid credentials before reaching the vulnerable parameter. The absence of UI:N and any scope-change prerequisite means no additional victim interaction or system compromise is needed once authentication is achieved.

The CVSS 4.0 vector AV:N/AC:L/AT:N/PR:L/UI:N with VC:L/VI:L/VA:L and threat metric E:P yields a score of 2.1, which is notably low for a network-exploitable SSRF with a public POC. The depressed score is driven by the low per-attribute impact ratings and absence of subsequent-system impact (SC:N/SI:N/SA:N). However, CVSS base metrics systematically undervalue SSRF in cloud-hosted environments where metadata services expose IAM credentials - a material escalation path not captured in this score. PR:L confirms that authentication is required, meaningfully limiting opportunistic mass exploitation compared to unauthenticated SSRF. No CISA KEV listing is present, so active widespread exploitation is not confirmed. EPSS data was not provided in the input; given the niche beta-stage product and public POC, real-world exploitation is plausible but likely targeted rather than opportunistic. The primary escalation concern is SSRF-to-cloud-metadata credential theft, which elevates practical risk beyond what the raw score implies.

An authenticated GoClaw user submits a crafted request to the invoke endpoint, setting output.video_url to an internal address such as http://169.254.169.254/latest/meta-data/iam/security-credentials/ targeting cloud instance metadata. The server fetches the URL without validation and may return the response body, potentially exposing IAM role credentials to the attacker. A public proof-of-concept demonstrating the SSRF behavior has been disclosed via GitHub issue #1199.