GitGalaxy Scanner

Version updated for https://github.com/squid-protocol/gitgalaxy to version v2.4.6.

  • This action is used across all versions by 0 repositories.

Action Type

This is a Composite action.

Go to the GitHub Marketplace to find the latest changes.

Action Summary

Gitgalaxy is a tool that automates the assessment of full repositories, identifying security risks and refactoring targets. It uses a custom regex/lexical structural-analysis engine to analyze code without requiring compilation and provides detailed reports and visualizations, helping developers prioritize tasks and improve their code quality.

What’s Changed

🚀 GitGalaxy v2.4.6: The Structural Hardening Update

We are incredibly excited to announce the release of GitGalaxy v2.4.6. This release represents a monumental milestone in the maturity of our core extraction engine. Over the past cycle, we initiated a comprehensive lockdown of our line-by-line regex parsing architecture across all supported languages.

GitGalaxy has officially transitioned from a “best-effort parser” to a production-grade, mathematically verified forensic security engine.

🛡️ The Hardening Effort: From 700 to 5,468 Tests

Our extraction engine was built for blistering speed (~35,000 LOC/s) and absolute ReDoS immunity, relying on strict horizontal (re.M) matching boundaries rather than slow, brittle AST parsers. But at scale, speed means nothing without precision.

When we started this initiative, we relied on a baseline of ~700 tests. As of today, the GitGalaxy extraction suite boasts a massive 5,468 highly-constrained test cases.

Through Epic #813 and Epic #518, we individually audited, isolated, and hardened the structural signatures—Dependencies, Functions, Classes, and Arguments—for all 54 supported languages.

🧪 Our Methodology: The Pathological Gauntlet

We didn’t just write “happy path” tests. We threw the kitchen sink at our regex engine to see where it would break. For every single core Phase 1 rule across all 54 languages, we enforced a strict multi-tier testing standard:

  1. Idiomatic Validation: Ensuring that standard, real-world code conventions (like Apex decorators, Python generics, or Ruby yields) are captured flawlessly.
  2. Negative & Ambiguity Testing: Ensuring the engine cleanly rejects lookalike tokens and doesn’t get confused when a language shares a keyword between two structural rules.
  3. Pathological Survival (The Red Team): We intentionally fed the engine adversarial, edge-case formats. We injected extreme vertical formatting, deep ReDoS-inducing generic nesting (e.g., List[List[...), and macro-like syntax traps to ensure our engine either parsed them safely or rejected them cleanly.

We systematically eliminated dangerous \s+ cross-line bridges that allowed logic to “leak” across lines, replacing them with strictly bounded horizontal whitespace [ \t]+.

⚖️ Absolute Truth via the Golden Crucible

To guarantee that these strict theoretical boundaries didn’t accidentally break real-world functionality, we utilized our crucible_check.py audit system. Every single regex tweak was cross-referenced against a “Golden Master” manifest of our entire test corpus.

We confirmed that our fixes caused zero unintended topological drift. The architectural mass, dependency maps, and structural boundaries of the real-world codebase remain flawlessly intact.

🌟 What This Means for the Future

By locking down the parser at this foundational level, we have eradicated the “Garbage In, Garbage Out” problem for our LLMs. The AI now receives mathematically precise code boundaries, un-polluted by typosquatting or false-positive function signatures.

GitGalaxy is now ready to perform advanced vulnerability graphing, logic-bomb detection, and threat analysis with absolute, verifiable confidence.