UE AngelScript Skills for Claude Code and Codex
Three skills for projects using Unreal Engine with Hazelight AngelScript.
| Skill | Use it for |
|---|---|
angelscript-in-unreal | Find local bindings, implement gameplay and verify script changes |
as-integration-tests | Design world, PIE and latent tests using the project's harness |
emmsui | Build developer panels with the EmmsUI plugin installed in the target project |
Install in Codex
Extract the ZIP from Releases into a new directory named ue-angelscript-distribution. From its parent, run:
codex plugin marketplace add ./ue-angelscript-distribution
codex plugin add ue-angelscript-skills-for-codex@ue-angelscript-skills-for-codex
Open a new project chat and ask Codex to use $angelscript-in-unreal to locate an existing binding. These skills work directly with local project sources. For live editor operations, optionally install Unreal Engine MCP Skills.
Install in Claude Code
Register the GitHub marketplace and install the plugin.
claude plugin marketplace add coquid/ue-angelscript-skills-plugin
claude plugin install ue-angelscript-skills@coquid-as-skills
In a new Claude Code session, enter /ue-angelscript-skills:angelscript-in-unreal. Both clients load the same skills/ directory. To install from the release ZIP, replace the repository name in the first command with the relative path to the extracted directory, such as ./ue-angelscript-distribution.
For GitHub installs, use the marketplace and plugin update actions in Claude Code’s /plugin menu.
Migrate from the combined plugin
If angelscript-unreal-engine-skills-for-codex@angelscript-unreal-engine-skills-for-codex is installed, remove that plugin and its enablement entry from any project-specific config before enabling the split plugins. Preserve your existing editor MCP connection. Install this package for scripting and the UE MCP package for editor control. The three scripting skill names carry over.
Development
With Python 3.11+ in a virtual environment, run:
python -m pip install -r requirements-dev.txt
python scripts/validate_plugin.py
python -m unittest discover -s tests -v
python scripts/package_plugin.py
The generated ZIP is in dist/. See the verification procedure, source and split baseline and MIT license.