KnowledgeBase depth
10 million open-source projects, 317,000 unique vulnerabilities, 2,650 licenses. The dataset is the practical asset. Programs underestimate how much SCA value comes from the data rather than the engine.
Black Duck • Application security
Black Duck SCA covers open-source dependency, license, and supply-chain risk against the Black Duck KnowledgeBase, which catalogs more than 10 million open-source projects, 317,000 unique vulnerabilities, and 2,650 unique licenses. Snippet matching detects copy-pasted and AI-generated code, and AI Model Risk Insights extends visibility into open-source AI model usage.
Merito sells Black Duck SCA and operates the policy authoring, snippet match tuning, SBOM lifecycle, and AI Model Risk Insights rollout that turn the scanner into a working supply-chain governance program.
What it is
Black Duck SCA is the Software Composition Analysis product backed by the Black Duck KnowledgeBase. The KnowledgeBase catalogs more than 10 million open-source projects, 317,000 unique vulnerabilities, and 2,650 unique licenses. The dataset depth is the practical reason Black Duck SCA produces useful findings on enterprise codebases where shallower datasets miss vulnerabilities.
Multiple scan technologies cover source code, build artifacts, containers, and firmware. Direct and transitive dependencies declared in package manifests get identified alongside dependencies the package manager does not declare (vendored code, copy-pasted snippets, AI-generated code that resembles open-source). Snippet matching is the differentiator that surfaces what manifest-based scanning misses, including AI-generated code from Copilot, Cursor, and Claude Code that resembles licensed open-source.
AI Model Risk Insights shipped in 2025 extending the SCA surface into open-source AI models. Visibility includes model usage, licensing, and data-origin information for the AI artifacts integrated into the customer's software. The capability addresses an emerging governance gap as enterprise programs adopt open-source AI models without the dependency-tracking discipline they apply to traditional libraries.
Black Duck SCA produces signal only when policy is authored. Default settings generate volume that gets ignored. Merito's standard SCA engagement begins with the dependency baseline, then policy authoring against the customer's risk tolerance and license tolerance, then SBOM lifecycle workflow, then audit-evidence flow. Programs that buy SCA and never write the policies generate noise that the security team learns to ignore.
Ideal use cases
What it is best at
10 million open-source projects, 317,000 unique vulnerabilities, 2,650 licenses. The dataset is the practical asset. Programs underestimate how much SCA value comes from the data rather than the engine.
Copy-pasted snippets and AI-generated code that resembles licensed open-source get matched back to source projects. Catches what manifest-based scanning misses entirely. The differentiator becomes more important as Copilot, Cursor, and Claude Code generate substantial portions of new code.
Source code, build artifacts, containers, and firmware all scan against the same KnowledgeBase. Single-surface SCA forces customers to stitch separate scanners; Black Duck covers the full footprint under one engine.
Open-source AI models get the same dependency-tracking discipline traditional libraries already get. Visibility covers usage, licensing, and data origins for the AI artifacts integrated into customer software.
Core capabilities
How Black Duck SCA finds the open-source code your application actually uses.
Manifest-based dependency identification
Direct and transitive dependencies declared in package managers (npm, Maven, NuGet, pip, gems, Go modules, Cargo, Composer) get identified and matched against the KnowledgeBase.
Build-artifact analysis
Post-build artifacts get scanned without source code access. Catches dependencies that exist in the deployed binary even when the package manifest is incomplete.
Snippet matching
Copy-pasted and AI-generated code gets matched against open-source projects. Catches what manifests cannot see, including code from Copilot, Cursor, or Claude Code that resembles licensed open-source.
Container and firmware scanning
Container images and firmware get scanned for embedded dependencies. Single-surface SCA misses the long tail of components shipped in deployed artifacts.
The proprietary dataset that drives finding quality.
10 million open-source projects
Project metadata, version history, and component relationships catalog the open-source ecosystem at depth that exceeds free vulnerability feeds.
317,000 unique vulnerabilities
Vulnerability records exceed the standard NVD dataset and include private vulnerability research and earlier disclosure than the public databases.
2,650 unique licenses
License catalog covers permissive, copyleft, dual-license, and unusual project-specific licenses. Drives accurate license-compliance enforcement.
Where SCA findings become enforced policy and audit-ready evidence.
Policy authoring
AppSec architects write policies covering vulnerability severity thresholds, license-compliance categories, snippet-match handling, and component age. Policies enforce automatically across the SDLC.
SBOM generation and import
Software Bill of Materials generated in SPDX and CycloneDX formats. SBOMs from upstream suppliers get imported and mapped to KnowledgeBase components.
Build-gate enforcement
Policy violations gate CI/CD builds. Block-on-introduce policy stops new vulnerabilities while triage works the legacy backlog.
Recent capabilities extending SCA into AI and software-supply-chain risk.
AI Model Risk Insights
Open-source AI models get the same dependency tracking traditional libraries get. Coverage includes model usage, licensing, and data-origin information.
Reachability with Black Duck Signal
Reachability analysis demotes findings in unreachable code paths. The backlog reflects what an attacker can exploit rather than every vulnerable line.
Coverity Static pairing
Joint Coverity SAST and Black Duck SCA workflows produce automated SBOM generation and unified findings management.
Where it fits in the stack
Deployment and implementation
Licensing and packaging
Black Duck SCA (on-prem)
Customer-hosted SCA platform with full KnowledgeBase access and custom policy authoring depth.
Best for: Programs with regulatory or operational requirements that mandate on-prem deployment.
Black Duck SCA through Polaris (SaaS)
SaaS SCA capabilities under the Polaris unified console alongside SAST, DAST, IaC, and secrets.
Best for: Programs adopting unified SaaS AppSec rather than running standalone scanners.
AI Model Risk Insights add-on
Extends SCA into open-source AI model dependency tracking, licensing, and data-origin visibility.
Best for: Programs adopting open-source AI models and needing AI governance discipline.
Merito services
Merito sells licenses and the delivery work around them. Pick the service that matches where you are in the lifecycle.
Installation or Polaris tenant setup, scan-source onboarding, KnowledgeBase calibration, and policy authoring.
Explore service02Supply-chain program scoping for SCA adoption alongside Snyk Open Source, Sonatype Lifecycle, Checkmarx SCA, and OpenText SCA.
Explore service03Build-gate SCA in Jenkins, GitHub Actions, GitLab CI, Azure DevOps, and Bitbucket. Container registry and package manager wiring.
Explore service04Developer-facing open-source approval workflows and AppSec champion programs.
Explore service05Named engineer, priority SLAs, and release-window coverage for Black Duck SCA in production.
Explore service06Long-term run support including ongoing policy tuning, snippet-match maintenance, SBOM workflow operations, and license-compliance review.
Explore service07Role-based training for AppSec architects, security engineers, and developers using Black Duck SCA findings.
Explore serviceBlack Duck SCA licensing
The KnowledgeBase is the asset. Default policy is the noise. Buy SCA through Merito and get the policy authoring, snippet-match tuning, and SBOM lifecycle together.
Merito point of view
Black Duck SCA's KnowledgeBase catalogs more than 10 million open-source projects, 317,000 unique vulnerabilities, and 2,650 unique licenses. The dataset depth is genuinely the differentiator. Programs that adopt SCA with shallow datasets miss vulnerabilities that Black Duck catches because the data is not in the free feeds. Merito's recommendation for customers comparing SCA tools is to weight dataset depth more heavily than engine architecture.
Snippet matching for AI-generated code is the recent capability that matters most for programs with significant Copilot, Cursor, or Claude Code adoption. AI assistants generate code that resembles licensed open-source, and manifest-based SCA cannot see it. Black Duck SCA matches snippets back to source projects so the license and vulnerability state of AI-generated code becomes part of the same backlog the rest of the SCA program manages.
AI Model Risk Insights extends governance to open-source AI models. The capability is necessary but not sufficient for AI governance. Customers that buy SCA for AI Model Risk Insights expecting it to cover model behavior monitoring are buying the wrong product. Merito treats AI Model Risk Insights as a dependency-discipline extension rather than a complete AI governance solution.
What buyers usually underestimate
Related from Merito
Related solutions
Related services
Related products
Frequently Asked Questions
Consultation request
Share your dependency surface, current SCA tooling, and supply-chain governance posture. A Merito Black Duck SCA specialist follows up within one business day.
KnowledgeBase depth
The dataset is the practical asset. Programs underestimate how much SCA value comes from data depth rather than engine architecture.
Snippet matching
Manifest-based SCA misses what package managers do not declare. Snippet matching catches the rest, including AI-generated code from Copilot, Cursor, and Claude Code.
Next step
A Black Duck SCA engagement with Merito starts with the dependency baseline, then policy authoring, then SBOM workflow. Default settings produce volume the security team learns to ignore.