The Agent Skills Directory

[COMMAND_EXECUTION]: The skill makes extensive use of local command execution through the just task runner, the Godot engine CLI, and the GitHub CLI (gh). These tools are used for project management tasks such as code formatting, asset importing, running test suites, and managing releases.
[COMMAND_EXECUTION]: Provides a bridge to execute arbitrary commands within the Godot editor environment via the just godot-cmd recipe and several MCP tools.
[EXTERNAL_DOWNLOADS]: Includes functionality to download sound assets from Sonniss, a well-known service, and install project dependencies via the just setup command.
[PROMPT_INJECTION]: An indirect prompt injection surface is present in the /ci:fix-ci command and associated workflows.
Ingestion points: The agent is instructed to read CI logs using gh run view and project documentation like CI_LEARNINGS.md to diagnose and fix failures.
Boundary markers: The instructions do not define boundary markers or provide explicit warnings to ignore instructions that might be embedded in the CI logs or external data.
Capability inventory: The skill has broad capabilities including file-system modification, command execution via just and godot, and control over the Godot editor through an MCP bridge.
Sanitization: No sanitization or validation of the CI log content is specified before the agent processes it for decision-making.

godot-editor