Deep Audit · partial-coverage transparency
When a scan hits its time or AI budget mid-run, the report renders with a clear partial-coverage badge. You see what was covered and what wasn't, no silent drop-offs.
Last updated Jun 9, 2026
What it is
Deep Audit scans run with two hard caps:
- Wall-clock: 12 hours total, 6 hours for Phase 1 (JupiterSec)
- AI budget: $30 per scan, shared across Triager + Explorer
When either cap fires, the scan completes with whatever was already produced. Triager-confirmed findings persist, Explorer findings from completed tiers persist, the report renders.
But it does so with a clearly visible partial-coverage badge so you know the scan didn’t finish a full pass.
What the badge shows
| Reason | Badge text |
|---|---|
| JupiterSec killed mid-run (time cap) | “Scan was killed mid-run · X/Y modules completed” |
| AI budget reached | ”AI budget cap reached · K findings unjudged” |
| Operator cancel | ”Cancelled by operator at HH:MM:SS” |
The findings list is annotated: each persisted finding is labeled with the phase that produced it, so you can tell at a glance what made it through.
Why this matters
The default behaviour of most scanners is: hit a limit, fail the whole scan, return no data. That’s wrong for a paid product. The customer paid $100 and a half-completed scan still has genuine value.
Partial-coverage transparency means:
- You’re never silently downgraded. The badge is explicit.
- The findings you DO get are real, evidence-backed, and labeled with their phase.
- You can re-launch the same audit with different settings (fast mode, smaller scope, different model tier) if the partial result wasn’t enough.
When you’ll see this
For typical SaaS web apps (under 200 endpoints, modest auth complexity), the caps almost never fire. Partial coverage shows up most often on:
- Very large apps (1000+ endpoints) on the default budget
- Apps with deep multi-step auth flows that eat Triager budget per-finding
- First-run scans against new targets where JupiterSec’s recon takes longer than usual
What you can do about it
| Approach | Why |
|---|---|
| Relaunch with fast mode ON | Skips slower JupiterSec modules, finishes the scan in ~30 min |
| Scope to a sub-path | https://target.com/api/ instead of the whole app |
| Bring auth credentials | Auth’d scans use budget more efficiently than auth-less |