Skip to content

Usage

REPL commands

Command Description
autopwn <ip> Full automated engagement against an authorized target
scan <ip> Deep port/service scan (-sV -sC, full range)
/run <action> Explicit single action, e.g. /run check smb vuln on 10.0.0.5
/ask <question> Ask the assistant (conversational; never runs tools)
authorize [ip\|cidr …] Show scope, or attest authorization for a scope
roe Show the active Rules of Engagement
findings / ports / vulns / creds / hashes / sessions Inspect results
verify Re-verify findings against the target
suggest Suggest next actions / access paths
plan Show the engagement plan (dry-run)
conclusion Classify and show the engagement outcome
save [path] Save findings as JSON
report <ip> Generate a PDF report
sessions / resume <id> List / resume saved engagement sessions
mode confirm\|auto\|never Set NL execution mode
verbose 0..3 Set verbosity
status / config / stats / history / clear Diagnostics & state
abort Abort the running engagement
exit / quit Leave the shell

authorize is available in tab-completion.

Typical session

hydra·sight › authorize 10.10.10.0/24
attest › I AUTHORIZE
hydra·sight › autopwn 10.10.10.15
hydra·sight › verify
hydra·sight › conclusion
hydra·sight │ report 10.10.10.15

Safety reminders

  • Plain conversational text never executes tools; use /run for explicit actions.
  • Every blocked or allowed command is recorded in hydrasight_audit.jsonl.
  • Nothing runs against a target outside the ROE ∩ authorization intersection.