60-second tour
For people evaluating the tool. Six commands, six outcomes. Run them in order; the next one assumes the previous one ran.
| Time | Command | What you should see |
|---|---|---|
| 0:05 | scripts/setup.sh --minimal | Python venv + adb + jadx + apktool + frida land under ~/.mnexus/. ~2 min. |
| 0:10 | source ~/.mnexus/env.sh && mnexus doctor | Engine matrix prints. The minimal engines should all be ● OK. |
| 0:20 | mnexus scan ./target.apk | Spinner runs, then a panel prints with the risk score + finding counts. ~30s on a 20 MB APK. |
| 0:35 | mnexus | The REPL opens. You’re in the cyberpunk terminal. Type /help if curious. |
| 0:45 | /serve followed by /open | The web UI launches in your browser. Sidebar lists every project tab. |
| 0:55 | /report html | ~/.mnexus/workspace/reports/PRJ-….html lands on disk. Open it. Mitigation Playbook is at the bottom. |
That’s the entire happy path. If any step fails the doctor explains why; everything else is incremental detail.
What to evaluate
- Did the risk score feel honest? Compare against your own gut on
the target APK. The scoring is tunable in
mnexus/intelligence/risk_score.py. - Did the Mitigation Playbook give you actual code? That’s the bar — if any finding shipped without remediation, that’s a bug, file it.
- Was the web UI fast? Static SPA, JSON over HTTP. 30 ms per route on a M-series Mac is the baseline.
- Did the auto-generated hooks compile when loaded into Frida? Open
the Hooks tab in the SPA and click Load to active session (or
copy the JS and feed it to
frida -U -l hooks/foo.js -f com.target).
If two of those four felt good, go to Workflows →. If none did, open an issue on the repo with the project id and what disappointed you — that’s the fastest feedback loop we have.