Purpose
Detects OS command injection using three strategies: blind OOB detection (reads callbacks from mg-oob), error-based detection via known error strings in responses, and time-based blind detection with sequential per-parameter baselining. Endpoints are sourced from the crawl corpus.
Output
cmdinject/results-<timestamp>.json— per-parameter findings with strategy, payload, and evidence.
CLI
mg-cmdinject acme-bounty
mg-cmdinject acme-bounty --oob-url https://oob.example.com/token