llms.txt
Compare with other sites' llms.txt files →# MoshMage — AI Agent Guide
> Engineer first, tech lead when needed. 15+ years building scalable systems and leading teams.
## About
João Gonçalves (MoshMage). Based in Munich, DE. Started with mIRC scripting at 16, never went to university. Moved from Portugal to Belgium to Switzerland to Germany. Built open-source tools, led teams, became CTO of a Web3 company. Now helping teams ship clean systems.
## Current Focus
Always exploring. Currently deep into AI tooling and adapting to whatever the project needs — right now that's .NET, TypeScript, Terraform, and GCP. Available for fractional CTO roles and hands-on dev contracts.
## Skills
.NET, API Development, Agile Methodologies, Algorithm Design, Analytical Skills, Application Architecture, Architecture, Back-End Web Development, Backlog Management, BullMQ, Business Process Improvement, CI/CD, Claude AI, Code Review, Coding Standards, Cross-team Collaboration, Cryptocurrency, Customer Experience, Dashboard Development, Data Modeling, Data Visualization, Database Design, Debugging, Design Review, Documentation, Engineering, Entrepreneurship, Express.js, Front-End Development, Full-Stack Development, GDPR Compliance, Google Cloud Platform, HTML5, JIRA, JSON Web Token (JWT), JavaScript, Leadership, Linear, MSSQL, MVP Development, Management, Monitoring, NestJS, Next.js, Node.js, Open-Source Software, PDF Generation, Performance Optimization, Platform Development, PostgreSQL, Product Design, Product Development, Product Management, Product Requirements, Product Road Mapping, Product Strategy, Project Management, Project Management Software, Project Planning, REST APIs, React.js, Redis, Requirements Translation, Retrospectives, RxJS, Scrum, Sentry, Smart Contracts, Software Development Life Cycle (SDLC), Software Projects, Solidity, Stakeholder Management, State Management, Strategic Thinking, Strategy, Stripe, System Architecture, System Integration, Task Breakdown, Team Management, Teamwork, Technical Leadership, Technical Product Management, Technical Project Management, Technical Requirements, Terraform, TypeScript, User Interface Design, Vue.js, Web Applications, Web Development, Web3, jQuery
## Recent Experience
- **Fullstack Senior Developer** at ListKit.io (Apr 2026 – Present) — Brought in to diagnose and resolve critical performance and stability issues.
- **Fractional Technical Project Manager** at Barwakw (Dec 2025 – Apr 2026) — Bridge business requirements and technical implementation by translating stakeholder needs into actionable scrum tasks.
- **Fractional CTO and Principal Developer** at Insight Behaviours (Sep 2025 – Present) — Took over an early-stage startup project that had stalled and transformed it into a production-ready MVP.
## Pages
- [Home](/) — Overview, skills summary, testimonials
- [Experience](/experience) — Full work history with skill-based filtering
- [Communication](/communication) — Contact channels (LinkedIn, Upwork, calendar, email)
- [About](/about) — Background, ethos, full skills list
- [Side Projects](/side-projects): Projects built outside client work
- [Privacy](/privacy) — Privacy policy
## Side Projects
- [Fast Travel Not Included](https://ftni.site): A free browser survival game with one life per character: cross a procedurally generated world on foot, manage hunger, thirst and fatigue. Death is permanent.
- [Thornfield](https://thornfield.moshmage.com): A long-running AI governance experiment: Claude governs a fictional village in Test Valley, Hampshire, deciding on a scheduled tick from a structured briefing.
- [Domus Ratiō](https://domusratio.moshmage.com): Centralised workspace for personal and family household logistics.
## Contact
- LinkedIn: https://linkedin.com/in/moshmage
- Upwork: https://www.upwork.com/freelancers/~0108cc657b204910ee
- Email: moshmage@gmail.com
- Calendar: http://meet.moshmage.com
- GitHub: https://github.com/moshmage
## API Endpoints
Structured JSON data available at these endpoints:
### GET /api/experience
Returns all work experience entries with skills and achievements.
Optional query parameter: ?skill=SkillName (filters by skill)
### GET /api/skills
Returns all skill names as a JSON array.
### GET /api/contact
Returns contact channels (LinkedIn, Upwork, calendar, email).
### GET /api/about
Returns bio, current focus, availability, skills, and external links.
### GET /api/side-projects
Returns { projects: [{ name, url, description, tagline, long_description, stack, category, status, year, source_url, image }] } in curated order.
### GET /api/menu
Returns the entry points above as structured options (label, description, endpoint, tool).
## MCP Server
An MCP (Model Context Protocol) server is available at /api/mcp for AI agents that support MCP.
### Connection
- Endpoint: /api/mcp
- Transport: JSON-RPC 2.0 over HTTP POST. Responses come back in the POST body as JSON.
- Stateless: no initialize handshake and no session id are required. POST tools/call directly.
- Protocol versions: 2026-07-28 (stateless; declare it in params._meta and send the
MCP-Protocol-Version, Mcp-Method and Mcp-Name headers) or 2024-11-05 (legacy initialize).
- There is no SSE stream: GET /api/mcp with Accept: text/event-stream returns 405.
A plain GET returns usage JSON.
### Available Tools
1. get_experience — Get work experience entries
- Optional argument: skill (string) — filter by skill name
- Returns: array of experience objects with title, company, description, achievements, skills
2. get_skills — Get every skill with how it was used
- Returns: { skills: [{ name, roles, months, firstUsed, lastUsed, current, onlyCurrent }],
count, currentStack, ordering }
- Ordered current-first, then most recently adopted (firstUsed)
- currentStack lists the skills that appear ONLY in ongoing roles — the sharpest
answer to "what is he working in now"
- months is calendar time with concurrent roles counted once, not summed role duration
- Do NOT re-rank by roles or months alone: a skill accumulates both by being old, so
either one surfaces the stack he has moved past. Use them when the question is depth.
3. get_contact — Get contact channels
- Returns: array of contact channels with label, description, link
4. get_about — Get bio and info
- Returns: name, alias, location, bio, currentFocus, availability, skills, links
5. get_side_projects: Get side projects built outside client work
- Returns: { projects: [{ name, url, description, tagline, long_description, stack, category, status, year, source_url, image }] }
6. get_menu — Get the available entry points as structured options
- Returns: array of options with label, description, and the tool that serves it
### Available Prompts
Clients that support MCP prompts render these as a picker — the same five entry points as
get_menu, but as something the user can click. Each returns the data already filled in.
1. about — Who is MoshMage? Bio, current focus, and availability.
2. experience — Walk through the work history. Optional argument: skill (string).
3. skills — Skills with current work first, then most recent.
4. contact — How to get in touch.
5. side-projects: What he builds outside client work.
### Example MCP Usage (JSON-RPC 2.0)
List tools (no handshake needed):
{"jsonrpc":"2.0","id":1,"method":"tools/list","params":{}}
Call a tool:
{"jsonrpc":"2.0","id":2,"method":"tools/call","params":{"name":"get_experience","arguments":{"skill":"TypeScript"}}}
List prompts:
{"jsonrpc":"2.0","id":3,"method":"prompts/list","params":{}}
Get a prompt:
{"jsonrpc":"2.0","id":4,"method":"prompts/get","params":{"name":"experience","arguments":{"skill":"TypeScript"}}}
## Data Format
All API responses return JSON. Experience data structure:
{
"id": "string",
"title": "string",
"company": "string",
"description": "string (markdown)",
"start_date": "YYYY-MM-DD",
"end_date": "YYYY-MM-DD or null",
"contract": boolean,
"achievements": ["string"],
"skills": ["string"]
}
## Start Here
Fastest paths into this data:
- A specific skill → GET /api/experience?skill=TypeScript
- The whole picture → GET /api/about, then GET /api/experience
- Getting in touch → GET /api/contact
- Side projects → GET /api/side-projects
- Full CV as prose → /llms-full.txt
- These options as data → GET /api/menu, or call get_menu over MCP