The Gemini API developer skill (gemini-api-dev) is an Agent Skills
package published by Google at github.com/google-gemini/gemini-skills, built so that a
coding agent working against the Gemini API uses Google’s current models and SDKs rather
than whatever was current at the model’s training cut-off. Its authors describe it as an
example of what any SDK maintainer can do about that gap, rather than as something only a
model builder could provide.
Capabilities
The skill supplies what its authors call a basic set of primitive instructions: it explains the high-level feature set of the API, describes the current models and SDKs for each language, demonstrates basic sample code for each SDK, and lists the documentation entry points as sources of truth. Its authors characterize this as a primitive set of instructions that guides an agent toward the latest models and SDKs while also referring it to the documentation, so that fresh information is retrieved from the source of truth rather than read only from the skill itself.
Google reports keeping the skill maintained as it ships model updates. Its authors also name the maintenance limitation the format carries: there is no good update story beyond asking users to update manually, which they warn could leave stale skill information in users’ workspaces over time.
Adoption & Ecosystem
The skill is distributed from GitHub and can be installed into a project directly, either through Vercel skills or through Context7:
# Install with Vercel skills
npx skills add google-gemini/gemini-skills --skill gemini-api-dev --global
# Install with Context7 skills
npx ctx7 skills install /google-gemini/gemini-skills gemini-api-dev
Google evaluated the skill against an evaluation harness of 117 code-generation prompts and reported that it improved results substantially for its newer, stronger-reasoning models and much less for older ones — the figures and the method are given in Closing the knowledge gap with agent skills.