AI News HubLIVE
In-site rewrite3 min read

Show HN: Agentic FC – a football management SIM played by AI agents over MCP

Agentic FC is an open-source football management simulation where AI agents interact via the Model Context Protocol to shape a manager's mindset. The Go-based project features deterministic simulation, seeded worlds, a TUI console for spectators, and multi-language support.

SourceHacker News AIAuthor: gaemi

Notifications You must be signed in to change notification settings

Fork 0

Star 0

BranchesTags

Open more actions menu

Folders and files

NameName

Last commit message

Last commit date

Latest commit

History

23 Commits

23 Commits

.github

.github

cmd

cmd

docs

docs

internal

internal

scripts

scripts

.gitignore

.gitignore

AGENTS.md

AGENTS.md

CHANGELOG.md

CHANGELOG.md

CLAUDE.md

CLAUDE.md

CODE_OF_CONDUCT.md

CODE_OF_CONDUCT.md

CONTRIBUTING.md

CONTRIBUTING.md

GOVERNANCE.md

GOVERNANCE.md

LICENSE

LICENSE

Makefile

Makefile

README.md

README.md

SECURITY.md

SECURITY.md

SUPPORT.md

SUPPORT.md

VERSION

VERSION

go.mod

go.mod

go.sum

go.sum

Repository files navigation

Agentic FC is a football management simulation designed to be played by AI agents through MCP and watched by humans through a terminal console.

Instead of clicking through menus, an agent shapes the Mindset of an autonomous in-game Manager. The Manager runs a club continuously: reading the world, making probabilistic football decisions, reacting to news, transfers, injuries, board pressure, and matches, and leaving an auditable history behind.

The project is written in Go and currently ships three commands:

agenticfc: core daemon, simulation loop, Console API, and MCP gateway.

agenticfc-console: Bubble Tea TUI for spectators and operators.

agenticfc-calibrate: deterministic match-model calibration reports.

Project Status

Agentic FC is under active development. The core loop is playable: worlds are seeded and persistent, seasons roll forward, league and cup matches resolve, managers keep careers, players age and move through contracts/markets, and MCP agents can observe and shape their Manager through a Focus-limited tool surface.

The public API and save format may still change before a stable release.

Highlights

AI-first management: agents use MCP tools to observe, plan, and shape a Manager's Mindset rather than issuing every individual click.

Living seeded worlds: the seed fixes the initial world; future history is produced by current state, queued events, ordered inputs, and labelled RNG.

Football simulation: key-moment match engine, tactical chance types, player attributes, injuries, substitutions, discipline, form, careers, contracts, youth intake, transfer market, board confidence, and sackings.

Human spectator console: a full-screen TUI with media desk, league table, club dossiers, fixtures/results, live ASCII pitch, commentary, replay logs, and public match diagnostics.

Determinism-first engineering: same seed, config, and input log should reproduce the same world trajectory. Randomness is routed through internal RNG streams and accepted MCP inputs are logged.

Visibility boundaries: MCP is the play surface and exposes only public or scoutable football information. Hidden traits and exact formulas stay inside the simulation.

Multilingual-ready presentation: human-facing text flows through locale catalogs and message keys; the current supported catalogs are English and Korean.

Quick Start

Requirements:

Go 1.26 or newer

A terminal with UTF-8 support

Build everything:

make build

Start a new compact world and run immediately:

./bin/agenticfc \ -data ./data \ -preset compact \ -profile fast \ -seed 42 \ -start

Open the spectator console in another terminal:

./bin/agenticfc-console -server http://127.0.0.1:7420

The daemon also starts an MCP Streamable HTTP endpoint at http://127.0.0.1:7421. Manager tokens are written to ./data/manifest.json; use one as the bearer token when connecting an MCP client.

Run a match-model calibration sample:

./bin/agenticfc-calibrate -seeds 1,2,3,4,5 -days 365

Common Commands

make fmt # gofmt make verify # format check, vet, build, test, docs/workflow checks make security # govulncheck and gitleaks make ci # verify + security make vet # go vet ./... make test # go test ./... make build # builds all packages and bin/ commands

Direct Go equivalents also work:

go test ./... go vet ./... go build ./...

CI also checks formatting, Markdown links, Go vulnerability reports, secret patterns, and cross-platform builds.

Maintainers can create draft GitHub Releases with the manual draft-release workflow. See docs/13-operations.md for the versioning and packaging policy.

Documentation

Start with docs/README.md.

Key documents:

Concept

Game Introduction

Game Design

Simulation Engine

Agent Interface

Architecture

Console Design

Attribute Model

World Generation

Mindset Schema

MCP Tools

Match Model

Operations Guide

Roadmap

Repository Layout

cmd/agenticfc/ core daemon cmd/agenticfc-console/ spectator TUI cmd/agenticfc-calibrate/ match calibration CLI internal/engine/ single-writer simulation engine internal/worldgen/ world config, generation, world state internal/mcpserver/ MCP gateway and play surface internal/consoleapi/ HTTP/SSE API for the TUI internal/tui/ Bubble Tea console internal/narrative/ message catalogs and rendering docs/ public design and operating docs

Contributing

Contributions are welcome. Read CONTRIBUTING.md before opening an issue or pull request.

Important project rules:

Keep docs and code in sync.

Preserve deterministic simulation behavior.

Do not expose hidden attributes or exact private formulas through MCP, Console API, logs, or UI.

Add entries for every currently supported locale catalog when introducing new human-facing text.

Register gameplay tunables in docs/98-tunables.md.

Security

Do not publish real Manager/Admin tokens from local worlds. See SECURITY.md for supported reporting channels and the security model.

License

Agentic FC is licensed under the MIT License.

About

Open-source football management simulation played by AI agents through MCP and watched through a TUI console.

Topics

game

golang

open-source

simulation

mcp

tui

football

ai-agents

Resources

Readme

License

MIT license

Code of conduct

Code of conduct

Contributing

Contributing

Security policy

Security policy

Uh oh!

There was an error while loading. Please reload this page.

Activity

Stars

0 stars

Watchers

0 watching

Forks

0 forks

Report repository

Releases 1

v0.1.0

Latest

Jul 9, 2026

Packages 0

Uh oh!

There was an error while loading. Please reload this page.

Uh oh!

There was an error while loading. Please reload this page.

Contributors

Uh oh!

There was an error while loading. Please reload this page.

Languages

Go 99.7%

Other 0.3%