Skip to content

// BENCHMARKS PÚBLICOS

Métricas de lab honestas, con caveats

Publicamos resultados reproducibles: fixture de regresión in-process (BENCH-01), probes multi-lab (BENCH-02) y runs en targets novel (GEN-01b). Las cifras son de runs reales, no objetivos de marketing.

Metodología completa y JSON en docs →

BENCH-01: fixture de regresión in-process (3 recall + 2 precision)

In-process regression fixture (3 recall + 2 precision cases). Starlette micro-lab — measures engine regression, not full-app recall on customer targets.

3/3
Recall (vulns encontradas)
2/2
Precisión (endpoints seguros)
0
Falsos positivos (CONFIRMED)
CasoFaseResultadoNotas
recall_cmdi_pingcmdiOKshell via host= (audit deep test)
recall_xss_searchxss_wafOKunescaped HTML reflection — WAF bypass phase
recall_ssti_greetsstiOKtemplate eval {{7777*7777}}
precision_safe_cmdicmdiOKhtml.escape — must not CONFIRM CmdI
precision_safe_xss_wafxss_wafOKhtml.escape — ENGINE-18 / audit H1

Targets novel (GEN-01b): motor limpio, sin seeds de lab

Scans autónomos surface-driven en apps que no tuneamos. Solo publicado tras un run real.

TargetRecall (core)Recall (extended)CoberturaNotas
OWASP NodeGoat3/38/9100%playwright_crawl OK (13 pages, 6 DOM forms) · 6 injection point(s) on bus · discover_api: 0 API endpoint(s) · sources={} · recall_core 3/3 on NodeGoat truth (`dast_detectable: yes`) — surface-driven motor (GEN-02): SSJI eval timing, profile write XSS, open redirect · no seed pack · truth matching is keyword/CWE heuristic

Probes multi-lab (BENCH-02)

Probes de recall en labs conocidos (jul 2026). Labs omitidos no cuentan como fallo. Scorecard completo: docs.

2/2 labs superados · 1 omitidos (lab no alcanzable)

LabResultadoHitsNotas
Broken CrystalsOK1/1GraphQL introspection / misconfig — SPA demo target
OWASP Juice ShopOMITIDO0/1REST search XSS recall on public demo instance
VAmPI (local)OK1/1BOLA on /users/v1/{username} — docker compose in demo/

Última ejecución: 7/2/2026, 7:50:04 PM