pwndeck · pen.protolabs.studio
protoPen runs the engagement itself — recon, WiFi, RF, RFID, OSINT — autonomously, on hardware you carry, against models you host. Headless-first, scope-aware, fully auditable.
____ _ _ ____ ___ __ | _ \ ___ ___| | __ | '_ \ \ /\ / / '_ \| | | |/ _ \/ __| |/ / | |_) \ V V /| | | | |_| | __/ (__| < | .__/ \_/\_/ |_| |_|____/ \___|\___|_|\_\ |_|
$ systemctl --user start protopen
[pwndeck] online · :7870 · engagement ready
$ goal "recon the lab subnet, report criticals"
Goal mode, wait / yield / resume, background sub-agents, and a monitor-goal cadence let it run unattended. A watchdog recovers it when a step stalls or the process dies. Set the objective and walk away.
Real radios, passed straight through a privileged container — monitor-mode WiFi with injection, software-defined radio capture/replay, and serial to the Flipper. One rig you carry, not a rack you ship.
The BlackArch repo ships configured, so any of ~2800 tools is one install away — and the agent already wraps the ones it reaches for: WiFi, recon, cracking, web, OSINT, telecom. Add your own skills, playbooks, and workflows on top.
Runs on your hardware against models you host. Engagement scoping and kill-chain phase-gating keep it inside the rules of engagement; goal verifiers are read-only and never execute shell. Every tool call lands in an immutable audit log.
Headless-first — but a full terminal-green console rides along: chat, an embedded PTY terminal, findings / targets / intel, and a semantic knowledge store. Drive it yourself, or chain it to other agents over A2A 1.0.
Not on a Deck? python -m server runs the agent + console
anywhere. Then wire the hardware via the
Steam Deck setup, or start with the
getting-started tutorial.
