Turn your best AI prompts into one-click tools in Chrome

My notes

Summary

Google launched Skills in Chrome: saved, reusable Gemini prompts that run one-click against the current page and any selected tabs. Activated via / or + inside Gemini in Chrome, plus a starter library of prebuilt Skills. Desktop-only, signed-in sync across Chrome devices.

Key Insight

• Prompts are now first-class browser objects, not transient chat turns. This is the same shift Raycast, ChatGPT custom GPTs and Alfred workflows made, now baked into the dominant browser.
• Multi-tab context is the killer feature: “select side-by-side specs across tabs” becomes trivial when you can pre-write the comparison prompt once and reuse against whatever tabs you have open.
• Confirmation gates for “sensitive actions” (calendar, email) suggest the agent track is where this is heading. Skills today is the training wheels for agentic Chrome tomorrow.
• A shared Skills library is a distribution channel Google controls. Expect prompt marketplaces, creator economy, and third-party Skill authoring down the line.
• Competitive pressure on standalone “prompt saver” tools (PromptBase, AIPRM, browser extensions). Native beats bolt-on for mass-market users.