Knowledge · Application Security

SAST Proof of Concept

A structured approach to conducting a SAST proof of concept, including codebase selection, ground truth creation, evaluation criteria, and decision-making methodology.

Primary question: How should organizations structure a SAST proof of concept to make an informed procurement decision?

Definitions

Proof of concept (PoC)

A time-limited, structured evaluation of a SAST tool against representative codebases and workflows, designed to generate empirical evidence about the tool's suitability for the organization.

Ground truth

A known set of vulnerabilities (both genuine and false positives) embedded in or identified within representative codebases, used as a reference for measuring SAST detection accuracy.

Seeded vulnerability

A deliberately introduced vulnerability in a codebase, used during SAST evaluation to measure detection rates and assess the tool's ability to find known issues.

The engineering problem

Organizations may run PoCs without clear criteria, ground truth, or measurable outcomes, resulting in inconclusive evaluations that do not support procurement decisions.

PoCs may test only against small or unrepresentative codebases, failing to reveal integration challenges or performance issues that would emerge at production scale.

PoC results may be influenced by vendor support during the evaluation period, which may not be available after procurement.

Security controls

Each control inspects a different artifact and produces evidence for an engineering decision.

Defined evaluation criteria

Clear criteria
Artifact
A documented set of evaluation criteria with assigned weights, covering detection accuracy, coverage, integration, analyst experience, and total cost of ownership.
Risk
Evaluating SAST tools against undefined or inconsistent criteria, making it impossible to compare results objectively.
Output
Defined evaluation criteria with clear scoring methodology.

Evidence:

Ground truth dataset

Known vulnerabilities
Artifact
A ground truth dataset of known vulnerabilities (both genuine and false positives) in representative codebases, used as a reference for measuring detection accuracy.
Risk
Inability to measure detection accuracy without a known reference.
Output
Ground truth dataset with documented vulnerabilities for accuracy measurement.

Evidence:

Representative codebase selection

Real-world testing
Artifact
A set of codebases that reflect the organization's typical development practices, technology stack, coding patterns, and architectural style.
Risk
Evaluating SAST tools against codebases that do not represent the organization's actual development practices.
Output
Representative codebases for realistic SAST evaluation.

Evidence:

Independent assessment

Vendor-independent evaluation
Artifact
A PoC assessment conducted independently of vendor influence, with results documented and scored objectively.
Risk
Vendor influence during the PoC period creating a biased or incomplete assessment.
Output
Independent assessment with objective results and scoring.

Evidence:

Verification workflow

  1. Define evaluation criteria and assign weights based on organizational priorities.
  2. Select representative codebases that reflect the organization's actual development practices.
  3. Create ground truth datasets with known vulnerabilities (both genuine and false positives).
  4. Seed additional vulnerabilities into codebases for controlled testing.
  5. Run each candidate SAST tool against the codebases.
  6. Measure detection accuracy against ground truth.
  7. Assess integration with existing CI/CD pipelines and security workflows.
  8. Evaluate analyst experience, remediation guidance quality, and reporting.
  9. Score each tool against defined criteria and document findings.
  10. Make a procurement decision based on empirical evidence.

Limits of verification

  • PoC results are time-limited and may not reflect long-term operational performance, including tool degradation, tuning requirements, or integration issues that emerge over time.
  • Ground truth datasets may not cover all vulnerability classes, frameworks, and code patterns present in production codebases.
  • Vendor support during the PoC period may not be available after procurement, potentially affecting the post-procurement experience.

Canonical terms used: SAST proof of concept; Proof of concept; Ground truth; Seeded vulnerability.

Evidence and references

  1. DerScanner SAST documentationDerScanner SAST analyzes supported source and binary formats, configuration files, and reporting and comparison of analysis results, with command-line interaction with CI systems and SSDLC integration.derscanner-sast

SAST evaluation

Structure your SAST PoC

DerScanner supports SAST proof of concept evaluations with broad language coverage and on-premises deployment.

SAST evaluation

Discuss SAST proof of concept

Share your current SAST evaluation process and challenges. We will help design an effective proof of concept.

Engineering knowledge for building and operating trustworthy systems.

DerSecur Recognition · build cad90ed · 2026-08-12 11:17:27Z · system