Release evidence
Shortcut Remover testing separates functional correctness, security-sensitive path handling, controlled-scale behavior, interface checks, and real drive-root observation. The goal is not to claim that one laboratory result predicts every PC; it is to make the release decision reproducible and its limitations visible.
Version 8.0.0 release gates
What the automated suite covers
- Shortcut parsing and risk-signal behavior.
- Path confinement and selected-root handling.
- Quarantine, restore, allow-list, and history workflows.
- Hidden-file visibility repair boundaries.
- Scheduling, cancellation, pause, and resume behavior.
- Localization resources and rendered interface states.
- Accessibility-oriented UI Automation and keyboard checks.
- Release artifact identity, manifests, hashes, and software-bill-of-materials generation.
Controlled shortcut corpus
The 8.0.0 release manifest records a labeled corpus containing 310 malicious test shortcuts and 600 benign test shortcuts. In that controlled corpus, all labeled malicious samples were detected and no labeled benign sample was selected as malicious. These are release-gate results for the included families—not a promise of perfect detection against unknown future threats.
Benign and malicious cases deliberately include some of the same Windows tools. This checks that the scanner considers command structure and context rather than treating the presence of PowerShell, Command Prompt, Rundll32, Regsvr32, or another administrative utility as an automatic verdict.
Scale and responsiveness gate
A controlled NTFS test root exercised 1, 1,000, 10,000, and 100,000 input-file scenarios. The larger scenarios injected deterministic 1% access failures and 1% malformed-shortcut parse failures so that error accounting could be checked rather than optimized away.
| Area | Observed 8.0.0 result | How to interpret it |
|---|---|---|
| 100,000-file scenario | Median 16,507 files/second; 95th-percentile elapsed 6.92 seconds | Controlled fixed-disk test data, not a promise for physical USB media. |
| Peak working set | About 71.4 MB across the gate | Observed in the release test environment. |
| Pause | No progress drift during the measured frozen window | Confirms workers stopped advancing during the controlled pause. |
| Cancellation | 95th-percentile response under 1 ms in the injected-delay harness | Measures cancellation signaling in the harness, not total device shutdown time. |
| Cleanup | No residual test files, directories, or harness workers | Checks that the gate leaves its controlled environment clean. |
Real drive-root verification
A separate read-only scan used the compiled release candidate against a real D:\ drive root. It completed naturally in 2 minutes 13.774 seconds and reported 42 directories, 17,486 files, 12 shortcuts, two access errors, zero reparse-point skips, and no safety-limit stop.
Because two locations could not be accessed, the application marked the scan incomplete rather than describing it as complete coverage. This test specifically verified the fix for an earlier defect in which a volume root could be normalized incorrectly and terminate almost immediately with zero files.
Independent review
The volume-root fix received a separate read-only audit of path containment, release artifacts, test output, installer contents, and the real drive-root scan. The review reported no P0, P1, or P2 findings; one best-effort statistics observation was recorded as non-blocking.
Known limits of this evidence
- The automated scale gate used a fixed local NTFS root; unattended physical USB hardware testing was deferred.
- Slow-media behavior was simulated with a cancellation-aware delay wrapper.
- Injected access and parse failures are deterministic and do not represent every real corruption pattern.
- Performance thresholds are release gates for the test machine class, not end-user guarantees.
- The current installer is not Authenticode-signed.
- Passing the release corpus does not guarantee detection of every future malicious shortcut.
How to verify your own scan
Check the selected path, elapsed time, directories, files, shortcuts, access errors, reparse-point skips, safety-limit status, and completion state together. A fast scan can be legitimate on a small or cached device; zero files or an unexplained incomplete state needs investigation. See why a full USB scan may finish too quickly.
Download the tested release View version 8.0.0 release notes