Fetch Archive to Lexiang

This skill will attempt to read and reuse your browser login state (Chrome cookie DB and/or connect to your real Chrome via CDP), may try to close and relaunch Chrome with debugging flags, and will persist session artifacts in your home directory (e.g. ~/.substack/storage_state.json and ~/.fetch_article/*). It also can call the OpenAI API for translations (OPENAI_API_KEY) and upload content to the Lexiang knowledgebase (requires a Lexiang/MCP token) even though those env vars/tokens are not declared in the registry metadata. Before installing: - Understand the risk: granting this skill access to your browser cookies or allowing it to start Chrome CDP effectively gives it the ability to fetch any content your browser can access (including paywalled content) and reuse your logins. - If you only want to fetch public pages, prefer running fetches in a disposable browser profile or run the scripts locally yourself rather than allowing autonomous agent actions. - Do NOT point the skill at your primary browser profile unless you trust the code and repository owner; instead create and use a dedicated Chrome profile or a separate VM/container. - Inspect the scripts yourself (they are included) and search for any hardcoded upload endpoints or unexpected network calls. Confirm where uploads go and how the Lexiang token is provided. - Provide OPENAI_API_KEY or Lexiang tokens only if you understand and accept how they will be used; expect these to be required for translation/upload features. - If you do install, consider disabling autonomous invocation (set disable-model-invocation = true in agent config) or run the skill with manual approval for each action. Given the sensitive local access and the metadata mismatches, proceed only if you are comfortable with these behaviors or can run the tool in an isolated/test environment.