oauth-sso-attack

Fail

Audited by Socket on Sep 23, 2026

3 alerts found:

Securityx2Malware
SecurityMEDIUM
SKILL.md

High-risk offensive skill. It is internally consistent with its stated purpose, but that purpose is to help an AI agent exploit OAuth/SSO flows, steal tokens/codes, and facilitate account takeover. No notable supply-chain concerns appear, but the exploit and exfiltration content makes the skill dangerous.

Confidence: 95%Severity: 93%
SecurityMEDIUM
evals/evals.json

This JSON artifact is not executable software but an instruction set designed to elicit and provide actionable OAuth SSO exploitation payloads and step-by-step account takeover workflows. It contains explicit malicious guidance (redirect_uri bypass patterns, missing-state CSRF/account binding flow, and OAuth/email auto-merge takeover). No obfuscation is present, and there are no direct malware execution behaviors within the file; however, the artifact has a high security risk due to its intended facilitation of real attacks when consumed by automated tools or shared without strict controls.

Confidence: 82%Severity: 92%
MalwareHIGH
references/oauth-attack-vectors.md

The provided fragment is highly indicative of malicious OAuth manipulation: it auto-submits an OAuth-related form using attacker-controlled grant/code/client/redirect parameters, including an attacker-controlled redirect_uri, consistent with authorization-code/token theft or account takeover mechanics. However, because no actual package/library source code was supplied, this cannot be used to attribute behavior to a specific open-source dependency implementation.

Confidence: 62%Severity: 98%
Audit Metadata
Analyzed At
Sep 23, 2026, 04:27 AM
Package URL
pkg:socket/skills-sh/wgpsec%2Faboutsecurity%2Foauth-sso-attack%2F@f12c5148bec5e81251fcbb97b0ae653b47405643679313098306a34fbb389468