AetherSAST-150-CWE-MCP news update news update

Maxxwwim released AetherScope V2, an architectural engineering environment and static application security testing cockpit combining real-time zero-day vulnerability scanning with a WebGL 2 visualization space. Detailed in github.com's report, the release packages a native Model Context Protocol server alongside cryptographic specification enforcement to eliminate context loss during automated artificial intelligence development. The software runs an integrated physics engine that models codebases as living spatial graphs alongside specialized static analysis tooling.

This report examines the core capabilities of the V2 release, including its documentation engine overhaul in version 2.2.2 and language additions in version 2.2.1. The system is designed for security auditors, systems programmers, and software architects managing complex codebases across multiple programming languages. Readers will find operational specifications for its static analysis rules, documentation parsing features, and dual-vision project classification engine.

01WebGL space cockpit and static scanning engine

The primary user interface operates an interactive two-dimensional and three-dimensional spatial graph powered by WebGL 2 signed distance field shaders. The graphics system maintains 120 frames per second using organic twelve-lobed cellular membranes, four mathematical layouts, and a Coulomb and Hooke physics simulation paired with procedural Web Audio. These visual components replace frozen architectural diagrams with dynamic spatial topology that models project relationships in real time.

Running underneath the visualization layer, the static analysis engine completes security audits in under 50 milliseconds upon every save command. The scanner evaluates source files against more than 185 zero-day patterns and Common Weakness Enumeration rules across Rust, PHP, Zig, C#, Python, Node.js, C, and C++. It features anti-false-positive lexical masking, typed vulnerability routing, and an interactive bug bounty scope and proof-of-concept generator.

Cryptographic governance anchors project contracts by enforcing rules for each folder through local markdown specifications. Every directory contract outlines Easy Approach to Requirements Syntax rules, line budgets, and import constraints sealed with BLAKE3 hashes in an aether lock file. Any architectural transgression detected by the engine triggers a red laser alarm within the visual space.

02Model context protocol server and agent gateway

Automated assistance connects directly through an integrated binary named aetherscope-mcp.exe. The server routes operational commands and state data using specialized handshake packets designed to bridge external machine learning assistants directly to the project structure. Compatible tools such as Antigravity, Claude, and Cursor communicate over this bridge without losing task context or hallucinating file locations.

Connected agents can automatically position the three-dimensional viewport to highlight targeted code bubbles across the graph. The bridge coordinates pending items inside a dedicated tasks directory while applying surgical file differences directly to the workspace. This mechanism prevents destructive file rewrites by isolating code edits to validated boundaries defined in the project architecture.

The gateway functions as an operational console that manages external access without granting unrestricted command-line access. By restricting modifications through structured protocol interfaces, the platform safeguards the active repository against unmonitored code generation errors and scope creep during multi-file refactoring sessions.

03Markdown documentation engine and interactive vector diagrams

Version 2.2.2 overhauls the internal documentation reader panel to deliver a self-contained reading environment matching standard web code repositories while remaining entirely operational offline. A dedicated tokenizer and parser pipeline ingests markdown and specification contracts, translating document elements into styled code blocks, metadata headers, and hierarchical navigation trees. Users can adjust the panel between standard 520-pixel, wide 880-pixel, and fullscreen 95-viewport-width layouts.

The documentation interface embeds the official Mermaid.js version 10 engine to render flowcharts, sequence diagrams, class structures, and state machines as scalable vector graphics. When running disconnected or prior to external script initialization, a built-in visual fallback tokenizer translates graph syntax directly into spatial cards. A dedicated diagram toolbar allows operators to copy source code, toggle between raw text and rendered graphics, or open full-definition zoom modals.

Code displays feature syntax highlighting across fourteen languages, including Rust, Python, JavaScript, TypeScript, Zig, C#, and SQL. The reader uses GitHub Dark theme palettes with an independent line-number gutter that isolates numerals from code selections during copy actions. Git diff blocks render line-by-line green additions, red deletions, and cyan metadata blocks alongside live preview panels for HTML and SVG snippets.

04Obsidian knowledge vault and dual-vision classification

Version 2.2.1 adds an Obsidian vault processing crate that automatically categorizes target directories into three structural modes. Pure vault mode activates when scanning directories containing Obsidian configuration folders or high densities of wikilinks without source code. Projects containing only programming languages trigger code architecture views, while hybrid directories enable both perspectives with dynamic graph switching.

The markdown parser resolves comprehensive knowledge-base syntax, including wikilinks, aliases, section anchors, block references, and transclusions. It calculates bidirectional link metrics to produce backlink counters and generates ghost note markers for uncreated references. File parsing captures inline tags and metadata frontmatter while filtering out common comments and code blocks from note indexes.

Directory traversal protection prevents arbitrary access attempts by validating file resolutions against strict boundary rules. The internal path guard rejects directory traversal sequences, server network paths, explicit operating system drive letters, null bytes, and external symbolic links. Users can toggle topologies between code and notes using dedicated view menus or keyboard shortcuts without reloading interface state.

05Systems language support and memory safety analysis

The release introduces native parsing and security checks for the Zig programming language through dedicated scanner and lexer modules. The scanner extracts structure definitions, packed types, external symbols, enumerated unions, error types, and import statements in parallel. The lexer applies lexical masking across multiline string literals while pruning test blocks from production analysis scopes.

Fifteen language-specific vulnerability rules evaluate Zig codebases for critical memory and pointer management defects. The analyzer flags use-after-free conditions under CWE-416, raw allocator resource leaks under CWE-401, and stack address returns under CWE-562. Additional rules identify uninitialized memory assignments matching CWE-457 and insecure pointer type conversions mapped to CWE-843.

The security scanning catalog extends across additional systems environments, incorporating static single assignment dataflow and reachability checks for software composition analysis. The analysis engine combines typed rule catalogs with community Semgrep signatures to provide comprehensive vulnerability identification across multi-tier enterprise projects.