← Back to Docs

vigil Documentation

Welcome to the documentation for vigil, the security scanner for AI-generated code.

Table of Contents

DocumentDescription
Quick StartInstallation, first scan, and basic concepts
CLI ReferenceAll commands, flags, and options available
Configuration.vigil.yaml file, strategies, overrides, and config merge
RulesComplete catalog of all 26 rules with vulnerable code examples
Output FormatsHuman, JSON, JUnit XML, and SARIF 2.1.0
CI/CD IntegrationGitHub Actions, GitLab CI, pre-commit hooks, and quality gates
DockerContainer usage, reference Dockerfile, and best practices
SecurityThreat model, what vigil detects, OWASP alignment, and limitations
AnalyzersTechnical reference for implemented analyzers (DependencyAnalyzer)
ArchitectureInternal structure, engine flow, analyzer protocol
Best PracticesRecommendations for teams using AI agents to generate code
ContributingGuide to contributing to the project, development setup, and testing

Project Status

vigil is in active development. The current version (v0.2.0) includes:

  • Full CLI with 5 subcommands (scan, deps, tests, init, rules)
  • Analysis engine with support for multiple analyzers
  • Dependency Analyzer active — detects hallucinated packages, typosquatting, nonexistent versions (DEP-001, DEP-002, DEP-003, DEP-005, DEP-007)
  • 26 rules defined across 4 categories
  • 4 output formats (human, JSON, JUnit XML, SARIF 2.1.0)
  • Configuration system with YAML, presets, and CLI overrides
  • 632 unit tests (~94% coverage)

The remaining analyzers (Auth, Secrets, Test Quality) are being implemented progressively.