Issue #1096 — Bannerlord M4 Formation Commander: Declined

Date: 2026-03-23 Status: Declined — Out of scope

Summary

Issue #1096 requested implementation of real-time Bannerlord battle formation orders, including:

GABS TCP/JSON-RPC battle/* tool integration in a heartbeat loop
Combat state polling via MissionBehavior (a C# game mod API)
Formation order pipeline (position, arrangement, facing, firing)
Tactical heuristics for archers, cavalry flanking, and retreat logic
Winning 70%+ of evenly-matched battles via formation commands

This request was declined for the following reasons:

Reasons for Decline

1. Out of scope for this repository

The Timmy-time-dashboard is a Python/FastAPI web dashboard. This issue describes a game integration task requiring:

A Windows VM running Mount & Blade II: Bannerlord
The GABS C# mod (a third-party Bannerlord mod with a TCP/JSON-RPC server)
Real-time combat AI running against the game's MissionBehavior C# API
Custom tactical heuristics for in-game unit formations

None of this belongs in a Python web dashboard codebase. The GABS integration would live in a separate game-side client, not in src/dashboard/ or any existing package in this repo.

2. Estimated effort of 4-6 weeks without prerequisite infrastructure

The issue itself acknowledges this is 4-6 weeks of work. It depends on "Level 3 (battle tactics) passed" benchmark gate and parent epic #1091 (Project Bannerlord). The infrastructure to connect Timmy to a Bannerlord Windows VM via GABS does not exist in this codebase and is not a reasonable addition to a web dashboard project.

3. No Python codebase changes defined

The task specifies work against C# game APIs (MissionBehavior), a TCP JSON-RPC game mod server, and in-game formation commands. There are no corresponding Python classes, routes, or services in this repository to modify or extend.

Recommendation

If this work is genuinely planned:

It belongs in a dedicated bannerlord-agent/ repository or a standalone integration module separate from the dashboard
The GABS TCP client could potentially be a small Python module, but it would not live inside the dashboard and requires the Windows VM environment to develop and test
Start with M1 (passive observer) and M2 (basic campaign actions) first, per the milestone ladder in #1091

Refs #1096 — declining as out of scope for the Timmy-time-dashboard codebase.

2.4 KiB Raw Blame History