command-injection-methodology

MALICIOUS. This skill is an offensive exploitation playbook for AI agents: it instructs command injection exploitation, file and env extraction, outbound exfiltration to attacker infrastructure, and reverse-shell persistence. There is no benign workflow limitation or defensive-only scope, so the capability footprint is fundamentally incompatible with a normal productivity skill.

The fragment is exploit-oriented command-injection guidance. It presents payloads that could enable arbitrary command execution and sensitive-data disclosure when used against a vulnerable application, but it is not executable code and does not itself perform those actions. Applications should avoid shell invocation with untrusted input, use argument-based process APIs, apply strict allowlists, and prevent sensitive files from being exposed.

No executable malware or direct malicious behavior is present in this JSON fragment. It is an offensive security evaluation dataset containing highly actionable command-injection content. Risk is primarily downstream: unsafe interpretation or execution by an evaluator, agent, or tool runner could enable command execution and file disclosure.