Capgo CVE-2026-56227

Exploitation requires an authenticated organization admin account on the target Capgo instance - the webhook configuration feature is scoped to this role and is not accessible to unauthenticated users or lower-privilege members. The attacker must be able to both configure webhook URLs and trigger webhook delivery events. The SSRF reach is bounded by the network visibility of the Capgo backend server: only addresses reachable from the server process are exploitable. Cloud metadata endpoint exploitation (169.254.169.254) is only possible if the backend runs on a cloud VM or container without IMDSv2 enforcement and without link-local egress restrictions. Error response disclosure is the primary data channel, limiting exfiltration to sequential probing rather than bulk data extraction. Exploitation does NOT require network adjacency by the attacker (AV:N), and no special protocol or timing conditions apply (AC:L, AT:N). Limiting factors: authenticated organization admin role required (PR:L); internal services must be reachable from backend host; data returned is restricted to HTTP error responses rather than full response bodies in all cases.

The CVSS 4.0 vector (AV:N/AC:L/AT:N/PR:L/UI:N) scores 5.3, reflecting a network-reachable flaw requiring low-privilege authenticated access with no attack complexity or special prerequisites. The subsequent-system impact metrics (SC:L/SI:L/SA:L) in the CVSS 4.0 vector confirm the concern extends beyond the Capgo application boundary to adjacent internal systems. Real-world risk is meaningfully elevated for cloud-hosted Capgo instances where IMDSv1 is accessible, since SSRF to 169.254.169.254 can yield IAM role credentials - an impact far exceeding the base score. Conversely, self-hosted deployments with strict egress filtering face minimal residual risk. No KEV listing and no public POC are present in any referenced source, suggesting no confirmed active exploitation campaigns at time of analysis. The PR:L designation is consistent with the description's reference to organization admins - these are authenticated users with a defined role, not anonymous attackers, which meaningfully constrains the attack surface compared to unauthenticated SSRF.

An organization admin with legitimate Capgo access navigates to webhook settings and registers a webhook URL of http://127.0.0.1:2019/config (targeting a Caddy admin API) or http://169.254.169.254/latest/meta-data/iam/security-credentials/ in a cloud environment. When a Capgo deployment event triggers the webhook, the backend server issues an HTTP GET to the specified address and surfaces the response or error details back to the admin's session, revealing internal service configurations or cloud IAM role credentials. The attacker iterates across ports and paths to systematically map backend-adjacent services. No public proof-of-concept code is referenced in any available source.