The Agent Skills Directory

[SAFE]: The skill is primarily instructional, offering detailed modding references and links for over 20 games and various modding frameworks (BepInEx, Harmony, Redscript, etc.).
[SAFE]: The skill provides two Python scripts, new_version.py and validate_project.py. Analysis of these scripts shows they are standard utility tools for managing an immutable semantic-versioning project layout. They use safe standard library modules (pathlib, shutil, re, argparse) and do not perform network requests, execute arbitrary commands, or access sensitive system paths.
[SAFE]: The skill implements a robust "Evidence Ladder" and "Non-negotiable laws" (found in MODDING-LAWS.md and GLOBAL-LESSONS.md) that explicitly instruct the AI agent against "inventing" API names, hook signatures, or asset paths. This is a strong proactive measure against hallucination-based errors that could lead to game crashes or instability.
[SAFE]: External URLs point to official game documentation (e.g., Paradox Wikis, Factorio API), well-known community wikis (e.g., BG3 Wiki, Noita Wiki), and established open-source repositories (e.g., BepInEx, tModLoader, CDPR-Modding-Documentation). These are standard and trusted resources within the modding community.
[SAFE]: The skill emphasizes a read-only approach to deployed game directories and mod-manager staging folders, ensuring that the AI agent only operates within a designated source workspace to prevent accidental corruption of the user's game installation.

game-modding