data-exfiltration-attack

Fail

Audited by Socket on Sep 19, 2026

4 alerts found:

Malwarex2Securityx2
MalwareHIGH
SKILL.md

SUSPICIOUS/HIGH-RISK skill. The install sources are mostly verifiable and same-project, but the skill’s core purpose is to equip an AI agent with offensive data-exfiltration, covert-channel, and DLP-bypass tradecraft. That makes the overall footprint dangerous and inappropriate for general agent use even though the documented capabilities align with the stated offensive purpose.

Confidence: 98%Severity: 95%
MalwareHIGH
payloads.md

This is an explicit data-exfiltration and command-and-control attack playbook. It contains numerous directly executable techniques for collecting files and cloud credentials, bypassing monitoring and DLP, transmitting data through covert or legitimate channels, establishing tunnels, executing commands, and erasing evidence. The red-team disclaimer does not materially reduce the capability or misuse risk. It is not obfuscated and is not an npm package implementation, but the content itself should be treated as high-risk offensive tooling and not executed outside a tightly controlled authorized lab.

Confidence: 99%Severity: 99%
SecurityMEDIUM
guides/data-exfiltration-attack-playbook.md

The fragment appears to be an offensive security/red-team exfiltration testing playbook rather than executable malware. It contains detailed, directly usable instructions for reconnaissance, covert data exfiltration, C2, credential retrieval, and DLP bypass. The material is potentially legitimate when used under explicit authorization, but it is inappropriate and dangerous if embedded in or executed by a general-purpose software package. No automatic malicious execution is demonstrated in the visible portion.

Confidence: 98%Severity: 93%
SecurityMEDIUM
test-cases.md

The fragment is not executable malware, but it is explicit offensive documentation for credential theft, covert command-and-control, internal pivoting, DLP bypass, and data exfiltration. If distributed in a general-purpose software package without a clearly authorized security-testing context, it represents a substantial supply-chain and misuse risk. No direct runtime malicious behavior can be confirmed from this documentation-only fragment.

Confidence: 99%Severity: 94%
Audit Metadata
Analyzed At
Sep 19, 2026, 04:39 PM
Package URL
pkg:socket/skills-sh/brucesongs%2Fkali-claw%2Fdata-exfiltration-attack%2F@fa5f7335b445a6c67d1b45f98cff5e30b1fd9d2c41475b3501d017e4abb204a0