W O L F E  ∞  J A M E S

Software Engineer & Format Architect - Creator of .FAF - Author of The 8 Eternal Cs

30 years designing formats that last. From Commodore Amiga to Internet AI standards, building technical infrastructure you can trust for the AI era.

Standards Body Work

AAIF Ambassador New

Founding Ambassador (2026 inaugural cohort) of the Agentic AI Foundation (AAIF) — the Linux Foundation project governing MCP, Goose, and AGENTS.md.

aaif.io • Onboarded June 2026

MCP Registry — Cargo Support

Authored and merged PR #1207 on the official Model Context Protocol Registry, adding Cargo (crates.io) as a first-class package registry type — schema, validator, and integration tests. Opened the documented registration path for the ~1,800 Rust MCP-related crates that previously had only a binary-packaging workaround. rust-faf-mcp was the first live, end-to-end validated example.

Merged June 3, 2026 • modelcontextprotocol/registry

IANA Media Type Registrations — 3

Registered three official Internet media types with IANA (Internet Assigned Numbers Authority) — the full FAF format family, each independently registered. The first, application/vnd.faf+yaml, is the first IANA-registered format with an official Anthropic MCP server.

vnd.faf+yaml (Context, Oct 2025) • vnd.fafm+yaml (Memory) • vnd.fafa+yaml (Agent identity)

IETF Internet-Drafts (Datatracker)

Two Internet-Drafts filed on the IETF Datatracker: the foundational .faf/.fafm context and memory format specification, and a follow-on draft defining .fafa — a portable identity format for AI agents.

draft-wolfe-faf-formatdraft-wolfe-faf-agent-01

CERN/Zenodo White Papers — 3

Three peer-citable academic preprints on Zenodo (CERN), each with its own DOI, covering the full FAF format family: context, memory, and agent identity.

.faf — Format-Driven AI Context Architecture
.fafm — Permanent Memory and Instant Recall
.fafa — Why Agents Need a Passport

W3C Community Groups

Member of the AI Agent Protocol Community Group, developing open, interoperable protocols that enable AI agents to discover, identify, and collaborate efficiently across the Web.

Also a member of the Autonomous Agents on the Web Community Group, focusing on Web-based multi-agent systems aligned with Web architecture, using Linked Data and Semantic Web standards for hybrid human-agent communities.

November 4, 2025 • Protocol standardization, multi-agent coordination, and Web architecture standards

Notable Projects

.faf Format (Foundational AI-context Format)

Universal format for AI project context. IANA-registered, validated by Anthropic (official MCP server), Google-approved (Chrome extension), and in production use across thousands of projects.

Downloads: 120,000+ across npm, PyPI, and crates.io
White Papers: 3 published on Zenodo (CERN), each with its own DOI
MCP Servers: 6 servers across Claude, Grok, Gemini, Cursor/IDE, and native Rust
Compilers: TypeScript, Rust (native + WASM), Zig (2.7KB WASM)

Visit faf.one →

The 8 Eternal C's of Computing

"Observed in 1995. Proven in 2025."

A universal framework describing all computation as an eternal cycle of eight operations: Context, Compile, Control, Cache, Compute, Communicate, Complete, and Cybernetics.

The theoretical foundation underlying the .faf architecture. Explains why Context-Mirroring works and how bi-directional sync enables persistent AI understanding.

Published: December 8, 2025
Status: Theoretical framework underlying .faf architecture

Interactive Overview → Read the Paper →

MCP Server Ecosystem (Claude, Grok, Gemini, Cursor/IDE)

First Anthropic-approved persistent project context server. Official MCP implementation merged into Anthropic's registry (PR #2759, October 2025). Expanded to 6 MCP servers across 4 AI platforms.

claude-faf-mcp: Official Anthropic MCP server, Desktop Extension (.mcpb)
faf-cli: Universal CLI (npm + Homebrew)
grok-faf-mcp: xAI Grok integration (npm)
gemini-faf-mcp: Google Gemini integration (PyPI)
faf-mcp: Cursor / IDEs / VS Code integration (npm)
rust-faf-mcp / faf-rust-sdk: Native Rust MCP server + SDK (crates.io)

Status: Production on 3 registries (npm, PyPI, crates.io)
Total Ecosystem Downloads: 120,000+
Architecture: stdio transport, Desktop Extension (.mcpb), native Rust + WASM compile targets

All Downloads → View on GitHub → View on crates.io →

AGENTS.md — NONE → GOOD → BETTER → BEST

A four-stage maturity model for agent instruction files: no AGENTS.md, any AGENTS.md, a facts-based AGENTS.md that stays true, and that discipline plus project.faf as the structured source of truth underneath it.

agents-md-facts: authors AGENTS.md from a repo's real facts — commands, entry points, conventions. Nothing invented.
mcp-better: applies the same discipline to MCP tool catalogs — a live host was found still running an outdated version against a newer published registry, a real example of the exact drift problem it exists to catch.

Featured by the Agentic AI Foundation on LinkedIn, August 2026 — "MCP mid-call is where 'honest tool catalog' and 'actually works' split."

agents-md-facts on GitHub → mcp-better on GitHub → View AAIF's post →

slash-tokens — Token Optimization for Context Engineers

Pre-flight token checks for LLM API calls. Know the cost before the call leaves your machine. 4.8KB WASM engine, sub-millisecond estimation, zero dependencies.

Features: Evaluate, route, abort — 90%+ tokens salvaged
Supports: Claude, GPT, Grok, Gemini
Downloads: 1,200+ (npm)

Visit slashtokens.com →

Historical Format Work (1995-2025)

Commodore Amiga .iff format (1994-1995): Early exposure to interchange file formats that shaped approach to format design.

Carpet Industry .prj format (1995): Created format for design specifications used by manufacturers across US, UK, Europe, India, Australia, and New Zealand. Still in use today.

CAD-SIM .sim format (2000-2016): Digital simulation format for US Carpet Industry. Deployed industry-wide at scale. Technology acquired by Shaw Industries (Berkshire Hathaway) in 2016.

HEX CODES - Hextra.io (2020-2025, ongoing): In-depth exploration of HEX/RGB/HSV-L color-space. Research and development in color format standardization and computational color theory.

AI R&D and Testing (2023-present): Extensive AI exploration and testing methodology development. Built WJTTC (WolfeJam Technical & Testing Center) - F1-inspired testing framework applying championship-grade standards to software development.

Contact & Credentials