Invoke Skill
AI-generatedSummary
Run Codex with one or more installed skills included in its instructions, using the supplied prompt as the task. Choose a model, reasoning effort, sandbox, working directory, and thread mode as needed; authentication comes from ProDex Setup on disk by default, or from n8n credentials when enabled.
Inputs
- Skills (required) — Select one or more installed skills to apply. Skills are loaded from the ProDex skills directory; selected skill instructions are added to the agent prompt.
- Prompt (required) — Instruction for Codex. Defaults to the incoming item's chatInput, prompt, or text field.
- Use n8n Credentials — Off by default: uses auth.json from ProDex Setup. Enable to authenticate with a selected ProDex Auth API credential.
- Model — Codex model: GPT-5.5, GPT-5.4 (default), or GPT-5.4 Mini.
- Reasoning Effort — Reasoning level: Extra High, High, Low, or Medium (default).
- Personality — Codex response style: Default, Friendly, or Pragmatic.
- Thread Mode — Start a new thread (default), continue the thread ID stored by this node, or resume a thread using a supplied ID.
- Thread ID — Thread to resume; shown when Thread Mode is Resume Thread By ID.
- Sandbox — Filesystem access for Codex: Read Only (default), Workspace Write, or Full Access.
- Working Directory — Optional directory Codex should use as its workspace.
- Options — Optional settings: Dynamic Skills accepts skill names, inline Markdown, or objects with name and content; Structured Output JSON Schema constrains output; Stream Progress To Execution Log logs progress; Timeout (Seconds) defaults to 300.
Output shape
One item per input item, each containing an agent result record.
The result includes output text, threadId, items, usage (or null when streaming), model, finishReason, and appliedSkills. Continuing a thread stores its threadId in node static data. Complete ProDex Setup before the first run and install the selected skills.