VAPT & penetration-test reporting

A pentest report that's never out of date.

The customer asks "do you have a recent pentest?" — TensorShield gives you a VAPT report you can hand over today, and again next month. Every finding grounded in real scanner evidence, mapped to CWE, OWASP Top 10 and MITRE ATT&CK, with a recommended fix and a signed attestation.

Markdown · JSON · signed evidence pack · regenerated on every scan

What's in the report

Every finding, fully worked.

It opens with a prose executive summary and an overall risk rating, then lists each vulnerability worst-first — and each one carries everything a security reviewer needs to act.

Severity + CVSS

Risk-rated and ordered worst-first, with the CVSS base score where the finding carries a CVE.

CWE + OWASP Top 10

Every finding maps to its CWE and its OWASP Top 10 (2021) category — the taxonomy an enterprise reviewer expects.

MITRE ATT&CK

Techniques attributed per finding, so the report speaks the language of a SOC and a red team.

Evidence strength — incl. captured PoC

Three tiers labelled inline: exploitation-proven (a reproducible proof-of-concept was captured), tool-confirmed (verified / corroborated), and pattern-match — plus a CISA KEV flag when it's actively exploited in the wild. The strongest tier carries the proof.

Recommended fix

An actionable remediation for the finding's class — and where TensorShield has already prepared the fix, it says so.

Tool & rule evidence

The exact scanner and rule that proves it. Nothing is asserted that a tool did not demonstrate.

Excerpt — one finding as it appears in the report
CriticalSQL injection in /search
Tool / rule: nuclei · nuclei::sqli-error-based
Location: https://app.acme.io/search?q=
CWE: CWE-89  ·  OWASP: A03:2021 Injection  ·  MITRE: T1190
CVSS: 9.8  ·  Evidence: verified · actively exploited (CISA KEV)
Recommended fix: Use parameterized queries / prepared statements; never concatenate user input into SQL. Run the app under a least-privilege database account. (TensorShield has already prepared this fix — it's awaiting your approval.)

Illustrative excerpt. Real reports contain only grounded findings from your own assets.

Quality & accuracy

At par with a manual pentest — by construction.

A report is only as good as its findings are true. TensorShield is built so the report can't claim what a tool didn't prove — the accuracy a good pentester delivers, made structural.

Grounded — never guessed

The engine can't record a vulnerability no tool supports. Every line in the report cites the scanner and rule that proves it (the anti-hallucination guard) — so there are no invented findings inflating the count.

Exploitation-proven, not pattern-only

The strongest findings carry a captured, reproducible proof-of-concept (exploitation-proven); others are labelled verified / corroborated vs pattern-match, and actively-exploited issues are flagged against CISA KEV — the accuracy signals a manual pentester earns by hand.

Standards-complete

CWE, OWASP Top 10 (2021) and MITRE ATT&CK on every finding, via the published crosswalks — the same taxonomy a $20k engagement deliverable uses.

Best-in-class detection underneath

The report is built on 30+ wrapped OSS scanners with recall on par with the standalone tools — depth that matches a human team's toolkit, run continuously.

Continuous vs one-off

A point-in-time pentest is stale the day after.

A traditional engagement is a snapshot — expensive, weeks of lead time, and out of date the moment you ship again. TensorShield runs the same assessment continuously and regenerates the report on demand.

TensorShield
continuous VAPT
One-off pentest
manual engagement
Coverage — web, API, code, containers, cloud, identity
Every finding grounded in tool evidence (no hallucinations)
CWE · OWASP Top 10 · MITRE ATT&CK mapped
Recommended fix per finding — and the fix shipped on approval
Signed, tamper-evident, reproducible evidence
Always current — regenerates as your stack changes
Turnaroundminutes2–6 weeks
Cost for an SMB$/mo$10–30k / test

A continuous automated assessment complements, and for many SMB needs replaces, a periodic manual engagement. For attestations that require a named human assessor, we can pair you with a partner — talk to sales.

How to get yours

From connect to report in minutes.

STEP 1

Connect your stack

Code, cloud, and identity over one-click OAuth — read-only by default. The agent discovers what to assess.

STEP 2

It runs the assessment

30+ scanners fan out across every asset, findings are verified and grounded, then mapped to CWE / OWASP / MITRE.

STEP 3

Download the report

A signed VAPT report in Markdown or JSON — regenerated on every scan, always reflecting your current posture.

Hand your next customer a pentest report.

Connect your first system and generate a signed, grounded VAPT report for free — then keep it current, automatically.