Aura

Product Introduction

1. Definition: Aura is an open-source semantic version control meta-layer (Apache 2.0) that integrates with Git repositories to track architectural logic via Abstract Syntax Tree (AST) hashing. It operates as a non-intrusive layer atop existing Git workflows.
2. Core Value Proposition: Aura solves traceability and safety gaps in AI-generated code by mathematically verifying logic integrity, enabling surgical rollbacks of broken functions, blocking undocumented AI commits, and reducing LLM token usage by 95%.

Main Features

1. Semantic AST Tracking: Hashes code structures (functions, classes, enums) instead of text lines using tree-sitter parsers. Creates immutable cryptographic identities for logic blocks, preventing "renaming death spirals" where Git loses traceability after refactors.
2. Amnesia Protocol: Enables surgical rewinds of specific functions/classes by swapping AST nodes. Removes hallucinated code from AI context memory to prevent recurrence. Uses differential AST analysis for precision.
3. Wave Execution: Orchestrates massive AI code generation in isolated, ephemeral environments. Prevents LLM context contamination by sequentially processing tasks in sandboxed contexts.
4. The Gatekeeper: Blocks forbidden logic (e.g., security flaws, proprietary leaks) via deep AST traversal. Projects architectural blast radius in real-time using dependency graphs.
5. Sovereign Vault: Dynamically redacts sensitive algorithms from contractor workspaces, replacing them with type-safe stubs. Maintains logic integrity while enforcing IP protection.

Problems Solved

1. Pain Point: Git’s line-based tracking fails for AI-generated code, causing unreadable diffs, lost logic context, and unchecked architectural drift during rapid AI iterations.
2. Target Audience: Engineering teams using AI coding assistants (e.g., GitHub Copilot), AI-native startups, enterprises scaling autonomous coding agents, and open-source maintainers.
3. Use Cases:
   • Block AI commits removing critical security checks (e.g., requireAuth()).
   • Roll back a single broken function in a 2,000-file refactor without merge conflicts.
   • Verify if AI-generated code violates architectural invariants before merging.
   • Audit 100+ AI commits/hour for compliance in regulated industries.

Unique Advantages

1. Differentiation: Unlike Git (text) or traditional SCM, Aura’s AST hashing provides mathematical proof of logic equivalence. Outperforms semantic competitors with 100% local operation and zero cloud dependencies.
2. Key Innovation: AST-based "Continuous DVR" captures every AI reasoning step at millisecond resolution, creating an immutable audit trail for hallucinations or compliance breaches.

Frequently Asked Questions (FAQ)

1. Does Aura replace Git?
   No. Aura is a meta-layer augmenting Git. Developers retain existing git commit/git push workflows while Aura adds AST-level traceability and AI safety checks.
2. How does Aura reduce LLM token usage by 95%?
   By isolating code generation contexts via Wave Execution, Aura eliminates redundant context reloading. AST-driven rewinds also minimize re-prompting for fixes.
3. Can Aura block emergency commits?
   Aura uses "warn and annotate" for architectural shifts but enforces hard blocks only for critical rules (e.g., plaintext secrets). Bypasses are allowed via --no-verify.
4. Is Aura suitable for non-AI workflows?
   Yes. Its AST tracking benefits any complex refactoring, but it’s optimized for AI-scale velocity (1,000+ commits/day) where Git fails.
5. How does AST hashing prevent false positives?
   Structural skeleton hashing ignores variable names/comments, focusing on logic topology. Mathematically equivalent functions yield identical hashes regardless of formatting.