Zennoxa Shield

Reproducible accuracy

OWASP Benchmark results

We run the industry-standard OWASP Benchmark and publish every number — including the categories where we still have work to do. Re-run it yourself with one command.

0.582
Benchmark score
92.5%
Precision
63.7%
Recall
5.5%
False positives
OWASP Benchmark 1.22,740 test casesMeasured 2026-08-04341 rulescommit 515fff7

Per-category breakdown

CategoryScoreRecallPrecisionFPR
Weak Randomness1.000100.0%100.0%0.0%
Insecure Cookie1.000100.0%100.0%0.0%
Weak Cryptography0.80880.8%100.0%0.0%
Weak Hashing0.69069.0%100.0%0.0%
Cross-Site Scripting0.51657.3%92.2%5.7%
LDAP Injection0.49959.3%84.2%9.4%
XPath Injection0.46746.7%100.0%0.0%
Command Injection0.45157.9%82.0%12.8%
Path Traversal0.44553.4%85.5%8.9%
SQL Injection0.30339.0%84.1%8.6%
Trust Boundary0.23747.0%79.6%23.3%

Score is Youden’s J (recall − false-positive rate), with strict OWASP CWE-per-category matching. Every category now registers detections; the lowest-scoring (trust boundary, SQL injection) are deliberately precision-first partial coverage, honestly reported rather than hidden.

Reproduce it

# clone the benchmark and run Shield against it git clone --depth 1 https://github.com/OWASP-Benchmark/BenchmarkJava && OWASP_BENCH=$PWD/BenchmarkJava make bench-owasp

Engine: SAST — regex + intra-file taint + dataflow (constant-fold / key-sensitive). Environment: 4 vCPU, go1.25.1, Linux. The archive is versioned and served at /api/benchmark.

OWASP Benchmark results — reproducible SAST accuracy — Zennoxa Shield