
from vigolium649
Prove OS command injection by generating OAST callbacks and a single safe follow-on data retrieval (file read or env dump) — non-destructive RCE proofing for se
This skill guides an agent through responsibly proving OS command injection vulnerabilities. It focuses on obtaining an OAST (Out-Of-Band) callback to prove execution and then running one non-destructive follow-on action (e.g., read a single file line or env dump) to size impact. The workflow prioritises safety and compliance: no destructive commands, no persistence.
Use this skill during penetration tests, DAST runs, or security audits when a parameter appears to reach a shell or child process (e.g., endpoints tied to ping, convert, ffmpeg, or user-supplied filenames). Trigger when metacharacters like ;, |, &&, backticks, or $() change responses, or when stack traces reveal execution binaries.
oast_mint and oast_poll, attack_kit payload patterns, and replay_request probes.Agents with network/OAST tooling and HTTP/DNS callback support (security scanning agents, DAST frameworks, or OpenClaw agents extended with oast utilities).
This skill has not been reviewed by our automated audit pipeline yet.