Skip to content

Software Intelligence: Version Vulnerability Intelligence Engine


Enter any software product name or common alias (e.g., Apache, Veeam, httpd, OpenSSL, LiteLLM) along with an exact version string to evaluate known applicable vulnerabilities and weaponized threat exposure:


The Hermes Software Intelligence engine follows a non-negotiable core architectural axiom:

Use deterministic structured data to determine vulnerability applicability; use intelligence to enrich, prioritize, correlate, and explain results.

In modern cybersecurity operations, relying on a large language model (LLM) to perform version comparison and interval testing produces hallucinations, dangerous false negatives, and false alarms. Hermes separates the stack into two independent tiers:

β”Œβ”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”
β”‚ DETERMINISTIC LAYER β”‚
β”‚ CPE / Product Resolution Β· Version Range Boundaries β”‚
β”‚ Exact Ingestion Math Β· CISA KEV Status Β· EPSS Scores β”‚
β””β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”¬β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”˜
β”‚
β–Ό
β”Œβ”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”
β”‚ INTELLIGENCE LAYER β”‚
β”‚ Exploit Maturity Β· Root Cause Analysis Β· Attack Paths β”‚
β”‚ Hermes Editorial Cross-Linking Β· Remediation Playbooks β”‚
β””β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”˜

A major flaw of legacy CVE aggregators is reducing complex configuration states to a binary Β« Vulnerable / Safe Β» judgment. Hermes strictly classifies every result into five standardized states:

Status IdentifierOperational Meaning & Criteria
AffectedThe supplied version falls strictly inside the documented affected version range without requiring optional non-default modules.
Configuration dependentThe version falls inside the affected range, but active exploitation requires specific non-default flags, compiled modules (e.g., mod_cgi), or architectural prerequisites.
Potentially affectedThe vendor advisory indicates theoretical exposure across major releases, but exact minor boundary verification remains incomplete.
Not affectedThe supplied version is strictly outside the affected boundary, or is equal to or greater than the official fixed version.
UnknownInsufficient structured CPE range data is available to make a deterministic decision. The system refuses to guess.

3. Fundamental Principle: β€œNo Known CVE” β‰  β€œSecure”

Section titled β€œ3. Fundamental Principle: β€œNo Known CVE” β‰  β€œSecure””

Hermes layers real-world exploitation telemetry directly over CVSS base metrics:

  • CISA KEV (Known Exploited Vulnerabilities): Flags whether threat actors, ransomware syndicates (e.g., Akira, Qilin), or initial access brokers are actively weaponizing the vulnerability in production environments.
  • EPSS (Exploit Prediction Scoring System): Quantifies the statistical probability (0.0% to 100.0%) that a software vulnerability will be exploited in the wild within the next 30 days.
  • Hermes Deep-Dive Correlation: Whenever a vulnerability is covered by original Hermes reverse-engineering research, a direct link connects the analyzer to our comprehensive disassembly, Sigma/KQL detection rules, and incident response playbooks.

The workbench features a Version Comparison Mode designed to evaluate upgrade paths (e.g., 2.4.49 $\rightarrow$ 2.4.51):

  • Eliminated Vulnerabilities: Flaws active in the source version that are completely neutralized in the target release.
  • Residual Risks: Vulnerabilities that persist even after upgrading, highlighting when a higher patch or secondary configuration hardening is necessary.
  • Regressions: Warnings if a target version inadvertently introduces newer unpatched CVEs.
  • Remediation Target Recommendation: The engine automatically analyzes available vendor patches to identify the lowest version that eliminates 100% of KEV and Critical threats.

The Hermes Threat Score (HTS) is a composite metric on a 0 - 100 scale that quantifies true operational exposure:

HTS = Min(100, (5 Γ— CVSS_max) + (15 Γ— EPSS_max) + (25 Γ— KEV_flag) + (10 Γ— Exploit_flag))
  • Base Severity (up to 50 pts): Anchored on the highest active CVSS score.
  • Weaponization Probability (up to 15 pts): Derived from the maximum EPSS percentile.
  • CISA KEV Weight (+25 pts): Applied immediately if any flaw is being actively exploited in the wild.
  • Public Exploit Availability (+10 pts): Added when verified weaponized exploit code or PoCs exist.

7. Asset Intelligence & SBOM Support (CycloneDX & SPDX)

Section titled β€œ7. Asset Intelligence & SBOM Support (CycloneDX & SPDX)”

The third workbench mode introduces client-side Software Bill of Materials (SBOM) ingestion, allowing organizations to audit multi-package software inventories in seconds:

  1. CycloneDX (JSON 1.4 β€” 1.6): Standardized specification designed by OWASP. The parser traverses root metadata components as well as nested dependencies, extracting names, exact versions, CPE identifiers, and Package URLs (purl).
  2. SPDX (JSON 2.2 β€” 2.3): ISO/IEC 5962 international standard. Packages are ingested from the packages collection, resolving security locators from externalRefs (cpe23Type, cpe22Type, purl).
β”Œβ”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”
β”‚ RAW SBOM MANIFEST (CycloneDX / SPDX) β”‚
β””β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”¬β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”˜
β”‚
β–Ό
β”Œβ”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”
β”‚ CANONICAL RESOLUTION PIPELINE β”‚
β”‚ 1. CPE Match: Identifies vendor & product identifiers β”‚
β”‚ 2. PURL Match: Parses package namespace & ecosystem types β”‚
β”‚ 3. Alias Resolver: Maps names via fuzzy canonical engine β”‚
β””β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”¬β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”˜
β”‚
β–Ό
β”Œβ”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”
β”‚ PARALLEL DETERMINISTIC AUDIT β”‚
β”‚ - Concurrent retrieval of partitioned product CVE datasets β”‚
β”‚ - Deterministic version-interval matching per component β”‚
β”‚ - Hermes Threat Score (HTS) calculation per component β”‚
β”‚ - Aggregate Inventory HTS & Exposure Gauge β”‚
β”‚ - Identification of Urgent Remediation Priorities β”‚
β””β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”˜

The overall inventory posture reflects both peak single-component exposure and horizontal multi-component risk:

Overall_HTS = Min(100, Max(HTS_components) + Min(15, (Vulnerable_Count - 1) Γ— 3))
  • Peak Component Weight: Anchored on the highest-threat asset in the fleet.
  • Cumulative Exposure Factor: Escalates the score by +3 pts for each additional vulnerable dependency (capped at +15 pts), accurately distinguishing an inventory with 1 vulnerable service from an environment harboring dozens of exposed entry points.
  • Actionable Remediation Playbook: Components harboring active CISA KEV exploitation or Critical severity CVEs are highlighted with recommended upgrade targets and instant 1-click links to run full Hermes Security Delta evaluations.

Every tracked core software component features a dedicated static security profile compiling overall posture, full affected version ranges, CISA KEV status, EPSS weaponization percentiles, and validated safe remediation targets: