Core
Every Core command in CodeFizz Editor Agent, with parameters and examples.
Every Core command in CodeFizz Editor Agent, with parameters and examples.
10 commands. Each shows its parameters and an example, and has a copyable deep link, so you (or an AI agent) can jump straight to one.
health_checkReport CLI, license, and editor-bridge status
Reports three independent layers and never fails just because no editor is open: (1) the CLI itself, (2) your license status (active/trial/expired, email, expiry) read offline from the shared store, and (3) the editor bridge - best-effort probe of the editor this folder targets, with a clear next step (open the project or 'cfa launch') when none is connected. 'healthy' reflects CLI + license only. Run this first to verify your setup.
cfa health_checkexecute_pythonExecute arbitrary Python code inside the running Unreal Editor
Runs Python code in the editor's Python environment where the 'unreal' module is pre-imported. Set a variable named 'result' to return structured data back. Use for operations not covered by other commands.
| Parameter | Type | Required | Description |
|---|---|---|---|
| code | string | Yes | Python code to execute (set 'result' variable to return data) |
cfa execute_python --code "result = str(unreal.EditorAssetLibrary.list_assets('/Game'))"dump_command_schemaDump the C++ reflection-generated schema for reflected commands
Returns the JSON Schema generated from each reflection-backed command's C++ param USTRUCT (the single source of truth). With --name, returns one command's schema; otherwise returns every registered reflected command's schema. Used to keep the Go CLI specs in sync with the C++ structs (see the schema-drift test).
| Parameter | Type | Required | Description |
|---|---|---|---|
| name | string | No | Optional: a single reflected command name to dump (omit for all) |
cfa dump_command_schema --name build_pcg_graphlist_editorsList every running Unreal editor (pid, project, engine, port)
Lists all Unreal editors currently running with the plugin: their pid, project name, full .uproject path, engine path, engine version, and bridge port. Use this to see what is running before connecting. Each command you run targets exactly one editor; use connect_editor to choose which.
cfa list_editorsconnect_editorSelect which running editor later commands talk to
Selects a running editor as the active target for subsequent commands (persisted, so it survives across separate terminals/processes). Choose by project name (e.g. --project jiggi, accepts 'jiggi' or 'jiggi.uproject') or by --port. Run list_editors first to see the choices. Use this when several editors are open, or when running from a folder that isn't inside the project.
| Parameter | Type | Required | Description |
|---|---|---|---|
| project | string | No | Project name to connect to (e.g. jiggi or jiggi.uproject) |
| port | int | No | Connect by exact bridge port instead of project name |
cfa connect jiggi
cfa connect_editor --project jiggi
cfa connect_editor --port 55557disconnect_editorClear the active editor selection
Clears the persisted editor selection set by connect_editor. After disconnecting, commands fall back to auto-detection (the current project's editor, or the single running editor) and refuse with a list when the choice is ambiguous.
cfa disconnectconnection_statusShow the active selection and which editor commands would hit
Shows the running editors, the current selection (if any), the current-folder project, and which editor a command would connect to right now (with the reason), or why it would refuse. Use this to understand routing before running other commands.
cfa connection_statusget_project_contextLive facts about this editor and project in one read
Engine version, platform and RHI, project name, path, engine association and open map, the compile-time capabilities of this plugin build (profiling, 5.7+, 5.8+), enabled plugins (or a yes/no check of the names you pass), and the asset inventory by class under a content path. get_skill_context uses it to fill the bundled skill.
| Parameter | Type | Required | Description |
|---|---|---|---|
| plugins | string | No | Comma-separated plugin names to check as enabled or not; empty lists every enabled plugin up to max_plugins |
| max_plugins | int | No | Cap on enabled plugin names listed when plugins is empty (0 = count only) |
| asset_path | string | No | Content path whose assets are counted by class |
| max_asset_classes | int | No | How many asset classes to report, most numerous first (0 skips the inventory) |
cfa get_project_context
cfa get_project_context --plugins IKRig,Water --max-asset-classes 10get_skill_contextThe bundled skill page with live facts from the running editor filled in
Reads one page of the bundled skill (SKILL.md or a reference page) and fills its {{tokens}} from the running editor in one round trip: engine version, project and open map, the plugins the page cares about, the commands unavailable on this engine version or without a plugin, the project's assets by class, and per-page extras such as the Niagara module list. With no editor running it returns the static page with the fallback text, and says why.
| Parameter | Type | Required | Description |
|---|---|---|---|
| page | string | No | SKILL (default) or a reference page name such as niagara, sequencer, retargeting |
cfa get_skill_context
cfa get_skill_context --page niagaralist_skill_templatesThe substitution tokens the bundled skill uses and which pages carry them
Lists every {{token}} get_skill_context can fill, what each one means, the static fallback text skill install writes, and the pages that use it.
cfa list_skill_templates