Release LedgerContinuous Protocol Evolution

Engineering Changelog

Chronological ledger of protocol standards, engine breakthroughs, CLI tooling releases, and agent infrastructure improvements from the Glintbase Labs team.

v3.0.0September 2026Major Release

ARS 3.0 Standard, Multi-Modal MCP Suite, Flight Simulator & Single Cloud Gateway

The definitive release of the Agent Readiness Standard (ARS 3.0) and the Glintbase Developer Platform. Delivers the 119-check open taxonomy, dynamic denominator scoring, multi-agent persona flight simulation, multi-modal in-chat SVG journey trees, dual-transport MCP 3.0, and the authoritative cloud gateway at https://scan.glintbase.dev/api/mcp.

• 119 Protocol Checks• 4 Operational Layers• 17 Production MCP Tools• 9 Bundled Agent Skills• Cloudflare Quick Tunnel• In-Chat Journey SVG• 0-Token Deterministic Mode

Key Improvements & Capabilities

Unified Cloud MCP Gateway (`https://scan.glintbase.dev/api/mcp`)

Strict single remote endpoint standard supporting standard JSON-RPC 2.0 and Server-Sent Events (SSE) with sliding-window IP rate limiting (60 req/min), permanent HTTP 308 alias redirection, and optional Bearer auth.

In-Chat Multi-Modal Flight Simulation

Simulation tools emit live inline Journey Tree SVG image blocks (`type: 'image'`) natively inside Claude and Cursor chats, accompanied by a zlib-deflated base64url state hash (`#data=`) linking directly to the web visual replay cockpit.

Ephemeral Cloudflare Quick Tunnel (`glintbase mcp --share`)

Zero-configuration tunnel spawning a secure, public HTTPS tunnel for local MCP servers in milliseconds, enabling remote team agents and container workers to interact with offline environments.

ARS 3.0 119-Check Taxonomy & Dynamic Denominator

Codified 119 discrete checks across 4 operational layers (Discovery, Access, Usability, Payments). Replaced rigid denominators with archetype-based baselines (85 for devtools/docs, 100 for ecommerce) and unpenalized bonus upside.

Autonomous Flight Simulator (`glintbase simulate`)

Dual-engine multi-turn execution harness with persona emulators for claude-code (200k), cursor (128k), and perplexity (zero-JS). Measures Dollar Tax burn ($3/M in, $15/M out) and Schema Friction Index.

Autonomous Code Doctor (`glintbase fix`) & PR Drift Shield (`glintbase ci`)

AST-driven auto-remediation engine synthesizing llms.txt, auth.md, robots.txt, and Anti-SPA 404 canaries with 1-click Git PR branching (`--branch`) and zero-drift GitHub Actions quality gates.

CLI v3.0 Executionbash
# Run complete ARS 3.0 audit with embedded Flight Simulator
npx @glintbase/cli audit https://api.stripe.com --simulate

# Simulate Claude Code persona with multi-modal SVG output
glintbase simulate https://api.example.com --agent claude-code -i "Create payment intent"

# Share local MCP server via Cloudflare Quick Tunnel
glintbase mcp --share

# Auto-generate missing agent surfaces and stage Git PR
glintbase fix --branch glintbase/agent-readiness
v2.5.0August 2026Engine Architecture

E2B Isolated Runtime Sandboxing & Counterfactual Retesting

Integrated isolated runtime code execution into the scanner engine, enabling real-time validation of documentation curl commands, SDK snippets, and counterfactual before-and-after retests.

• E2B Linux MicroVMs• AST In-Memory Evaluation• Zero-Disk Blast Radius

Key Improvements & Capabilities

Runtime Sandboxed Execution

Integrated @e2b/code-interpreter to spin up ephemeral, secure Linux sandboxes that execute code examples discovered in docs to detect broken parameters and 500 errors.

Counterfactual 'What-If' Proof Engine

In-memory ArsSandbox automatically mounts suggested fixes to simulate agent re-runs, empirically calculating exact token savings and latency reductions prior to disk writes.

v2.2.0August 2026Engine Architecture

3D Force-Directed Knowledge Graph & Schema Friction Index

Introduced 3D topological visualization of agent-facing software surfaces and the Schema Friction Index for OpenAPI and MCP tool inputs.

• WebGL / Three.js 3D Graph• Zod/Ajv Parameter Analyzer

Key Improvements & Capabilities

3D Force-Directed Topological Graph

Powered by Three.js and WebGL, visualizing inter-page cross-references, API endpoint dependencies, and orphaned documentation clusters.

Schema Friction Index

Measures ambiguity in tool parameters, type mismatches between OpenAPI descriptions and schemas, and unit omissions that trigger agent hallucination loops.

v2.0.0August 2026Protocol Standard

WebMCP In-Browser Bridge & Anti-SPA Canary Detection

Launched the browser-based WebMCP provider allowing autonomous AI web agents to interact directly with web interfaces, alongside active canary probing for single-page app soft-200 failures.

• WebMCP Standard v0.9.1• Soft-200 Canary Probes

Key Improvements & Capabilities

WebMCP Standard Bridge

Exposes structured platform state, page navigation, and callable tools directly via window.modelContext for visiting AI browser agents.

Anti-SPA 404 Canary Probing

Active HTTP probes send randomized canary requests to detect SPAs that return 200 OK HTML shells for missing routes—the #1 cause of agent context pollution.

v1.5.0July 2026CLI Tooling

Shareable Scorecards & Competitive Leaderboard Benchmarks

Added dynamic 1200x630 scorecard image rendering for sharing scan results on social channels, and launched the public Agent Readiness leaderboard.

• 1200x630 Social Previews• Public Leaderboard Index

Key Improvements & Capabilities

Dynamic Social Scorecards

Automated server-side generation of high-resolution summary cards showing score band, layer breakdown, and verification badges.

Public ARS Leaderboard

Ranks hundreds of public software products by Agent Readiness Score, driving transparency in machine operability.

v1.0.0May 2026Major Release

Initial Launch: ARS 1.0 Composite & Deep Extraction Engine

The initial foundation of Glintbase. Established the original ARS 1.0 composite score, deep crawl extraction of JS-rendered frameworks, and preliminary agent journey simulation.

• 8 Core Dimensions• Embedded Data Extractor

Key Improvements & Capabilities

Deep Crawl Extraction

Extracted structured content from __NEXT_DATA__, RSC flight chunks, and JSON-LD from JavaScript-heavy documentation shells.

Preliminary Agent Journeys

Simulated whether language models could complete multi-step setup tasks using only documentation text.