Open WebUI

Complete Open WebUI API integration for managing LLM models, chat completions, Ollama proxy operations, file uploads, knowledge bases (RAG), image generation, audio processing, and pipelines. Use this skill when interacting with Open WebUI instances via REST API - listing models, chatting with LLMs, uploading files for RAG, managing knowledge collections, or executing Ollama commands through the Open WebUI proxy. Requires OPENWEBUI_URL and OPENWEBUI_TOKEN environment variables or explicit parameters.

This skill appears to implement the Open WebUI API as described, but the registry metadata incorrectly omits that an API token (OPENWEBUI_TOKEN) is required — the included CLI will fail without it. Before installing: 1) Inspect the provided scripts locally (openwebui-cli.py is included) to verify behavior. 2) Only use an API token for an Open WebUI instance you trust; uploads and chat messages (and any file contents) are sent to the configured OPENWEBUI_URL and could expose sensitive data. 3) Confirm destructive actions (deleting models/collections) only after explicit prompts. 4) Note the script defines a token-redaction helper but review any output/logging to ensure full tokens are not printed. If the registry metadata were corrected to declare the required env vars, and you are confident in the target Open WebUI instance, this would be coherent; the current mismatch is why I rate it suspicious.

Type: OpenClaw Skill Name: open-webui Version: 1.0.0 The skill bundle provides a legitimate API client for Open WebUI, implementing features like model listing, chat completions, file uploads for RAG, and Ollama proxy operations. The `SKILL.md` clearly defines the skill's purpose, prerequisites, and includes explicit safety instructions for the AI agent, such as never logging API tokens and requiring confirmation for sensitive actions like deleting models or accessing files outside the workspace. The `scripts/openwebui-cli.py` Python script adheres to these instructions, using standard libraries for network communication and file operations, and includes confirmation prompts for potentially destructive actions. There is no evidence of data exfiltration, malicious execution, persistence mechanisms, or prompt injection attempts against the agent.