User Scanner
User Scanner Performance and Cost: Benchmark Coverage, Pivots, and Review
A safe benchmark plan for User Scanner concurrency, module latency, proxy overhead, cross-scan breadth, report size, and analyst verification time.
What you will learn
- Explain module, pivot, report, and MCP layers
- Run a bounded scan and interpret uncertainty
- Apply authorization, rate-limit, and data-retention controls
Before you start
- Basic Git and command-line usage
- Comfort reading a project README
You can explain the project, run its documented first step, and decide what to verify before adopting it.
Key takeaways
- Benchmark complete bounded cases with local/owned fixtures.
- Report latency, coverage, blocks, bytes, proxy overhead, report size, and analyst review.
- Optimize accepted evidence quality and cost, not raw request count.
Measure a case, not just requests
The useful unit is a bounded authorized investigation: target normalization, module discovery, requests, pivots, report generation, and human verification. Use local mocks or owned targets, and label live network work separately.
Pin module set, fixture responses, concurrency, proxy mode, TLS adapter, Python version, package commit, and output format. Cold index refresh and warm scans answer different questions.
Separate latency and coverage
Record p50/p95 per module, queue depth, connection reuse, timeout/retry rate, proxy health, response bytes, CPU/memory, and report size. For cross-scan, measure breadth, depth, duplicate suppression, and critical-path time rather than only total requests.
Higher concurrency may improve throughput while increasing blocks, provider load, and review volume. Treat coverage and evidence quality as separate dimensions.
Price data and analyst time
Include proxy or API costs, storage, CI minutes, report review, false-positive verification, and deletion work. Calculate cost per accepted, defensible finding or exposure check—not per HTTP call.
A faster scan that collects too much or creates ambiguous pivots is not an optimization. Keep the same rubric and fixture when comparing a candidate change.
Publish a decision-ready report
Store raw timings, fixture hash, module snapshot, environment, redacted outputs, and reviewer notes. Set budgets for concurrency, request volume, storage, review minutes, and proxy spend; canary changes before broad use.
Decision guide
| Criterion | Option A | Option B |
|---|---|---|
| Best when | You need predictable behavior and easy auditing | You need adaptive optimization and have reliable telemetry |
| Main risk | May leave performance on the table | Can become difficult to explain or debug |
Implementation steps
- 1
Freeze fixture, modules, runtime, concurrency, proxy, and output.
- 2
Measure cold/warm, per-module, pivot, export, and failure paths.
- 3
Calculate resources, bytes, spend, review, and deletion cost.
- 4
Publish raw receipts, budgets, canary results, and rollback criteria.
Copy-ready example
fixture -> module p50/p95 + blocks/retries
-> bounded pivots -> report bytes/review minutes
-> cost per accepted exposure checkFrequently asked questions
Is more concurrency always better?
No. It can increase blocks, provider load, false positives, and review volume; benchmark the complete case.
Should I publish live target timings?
Only with authorization and careful redaction; local mocks are preferred for reproducible articles.
Sources
- User Scanner README (captured 2026-08-31)Source checked 2026-08-31
- User Scanner repositorySource checked 2026-08-31