Building a Zero-Dependency Python Parser to Convert Veracode SAST JSON to HTML Dashboards

In enterprise DevSecOps pipelines, velocity is everything. While running static application security testing (SAST) tools like Veracode inside containerized CI/CD runners is crucial for catching vulnerabilities early, dealing with the raw artifact data can be a bottleneck.

By default, high-velocity pipeline scans output a heavy, nested results.json file. Expecting developers or QA leads to dig through thousands of lines of raw JSON during a broken build step slows down remediation cycles.

To solve this, I built a lightweight, zero-dependency Python automation script that parses raw Veracode JSON data and instantly outputs a beautifully styled, responsive Bootstrap 5 HTML reporting dashboard.

📊 Full Engineering Architecture & Implementation Details: > For the complete step-by-step framework setup, local directory structures, and advanced break-build pipeline rules, check out the full guide on my portfolio:

Veracode SAST Pipeline Scan Automation Guide

---

🚀 Why Go Zero-Dependency?

When writing automated tasks for ephemeral CI/CD runners (like GitHub Actions, GitLab CI, or Bitbucket Pipelines), minimizing container setup time is critical.

Using external libraries like pandas or requests requires a pip install phase. This consumes precious build minutes, requires internet access within isolated runners, and introduces third-party dependency vulnerabilities.

This parser uses nothing but Python's built-in json and os libraries, meaning it executes in milliseconds on any minimal container base (like Python-slim or Alpine).

🛠️ The Core Logic Blueprint

The parsing script executes three main tasks:

1. Validation & Ingestion: Safely checks for the existence of the compilation artifact (results.json).
2. Metrics Aggregation: Tallies up total vulnerabilities and segments them into priority levels (High/Critical vs. Medium vs. Low) based on Veracode's severity weights.
3. Semantic HTML Synthesis: Flushes the parsed findings arrays into an optimized, self-contained Bootstrap 5 table component.

python
# Quick snippet showing the data extraction layer
total_flaws = len(findings)
high_flaws = sum(1 for f in findings if f.get("severity") >= 4)
med_flaws = sum(1 for f in findings if f.get("severity") == 3)
low_flaws = sum(1 for f in findings if f.get("severity") <= 2)