Google replaces Gems with Skills in Gemini chat
Google is replacing its Gemini Gems feature with "Skills," a move that aligns with OpenAI and Anthropic to transition toward standardized, agent-ready prompt formats.

Google has begun a global rollout of "Skills" in Gemini chat, replacing its previous "Gems" feature. This transition mirrors similar moves by OpenAI, which is sunsetting its Custom GPTs, and Anthropic, which originally created the open-standard Skills format. The shift represents an industry-wide alignment toward standardized, agent-ready prompt structures that simplify how users interact with customized AI behaviors.
Users can save frequently used instructions as a Skill and trigger them instantly by typing a slash followed by the Skill name. Gemini can also automatically generate Skills based on past chat histories and execute them when it recognizes a matching prompt. For more complex workflows, practitioners can chain multiple Skills together, such as combining a specific writing style with corporate brand guidelines.
Currently, Skills support reference materials in the form of text documents, PDFs, and images. Google plans to expand these capabilities in the coming weeks by adding sharing features, Google Drive file integration, and Gemini Notebook support. This will be followed by a broader rollout to enterprise, education, and nonprofit Workspace customers.
The transition timeline varies by account type. Gems will shut down in November for personal accounts, followed by March 2027 for enterprise and nonprofit Workspace users, and June 2027 for education customers. Google will automatically convert existing Gems into the new Skills format. Additionally, November marks the end of Opal, Google's AI mini-app experiment that debuted in the summer of 2025.
For AI practitioners and developers, this shift simplifies prompt engineering by moving away from isolated custom chatbots toward modular, interoperable prompt templates. The ability to chain Skills and trigger them via simple slash commands allows for more fluid, programmatic workflows within the standard chat interface, paving the way for more sophisticated agentic automation.
This is our own summary of reporting by The Decoder



