Scope-gated recon
Each runner checks the target against the sealed scope before it executes. Out of scope is refused without a packet sent. Discovered hosts become new targets, re-checked at every hop.
Operator console · mock engagement
This is the Wedjat engine walking a fictional in-scope estate. A Suite of recon runners fires in sequence, each one scope-gated before it touches anything. Findings stream into the console as they land. Open any one to see its proof: the exact command, the evidence hash, and the verdict. Find it, prove it, hand it off.
What you just watched
The engine does not hand you a risk score and a guess. It runs deterministic recon runners, records the exact argv and a hash of what it saw, and asserts a finding only when it can show the proof. The same evidence is what a report is built from, and what you can re-run yourself.
Each runner checks the target against the sealed scope before it executes. Out of scope is refused without a packet sent. Discovered hosts become new targets, re-checked at every hop.
A finding carries the real argv, a sha256 of exactly what was stored, the tool version, and the box image. A replayer can verify it. No evidence reference, no finding.
To show a door opens we write our own marker and read it back, then drop it. We demonstrate access. We never read your data, never flood, never exploit. Detection only.