Evidence record · captured 2026-10-11

Gemini API Development

Overrides the agent's training data with a current Gemini model list and critical rules, then guides SDK usage for text generation, chat, multimodal input, image/video/speech generation, streaming, function calling, structured output and long-running research tasks. A reference covers migrating from the older generateContent API.

Official publisherUnverifiedSource reviewedBehavior not tested

Source review

Editorial assessment

This is a source-only assessment. It does not claim runtime or security testing.

What it does

Tells the agent its training data is outdated, pins a current Gemini model and SDK list, and requires fetching the matching ai.google.dev page before coding, with examples built on the Interactions API.

Best for

Writing new Python or TypeScript code against the Gemini Interactions API
Moving existing generateContent code or older model strings to current Gemini models
Starting Deep Research or managed-agent jobs that run in the background and are polled for results

Strengths in the source

Requires fetching the relevant ai.google.dev .md.txt documentation page before writing code, and treats the inline examples as minimal starting points.
Flags easy-to-miss behavior: interactions are stored by default, and tools, system_instruction and generation_config must be re-sent on every turn.
The migration reference has the agent confirm scope before editing, sizing it with an rg count of generateContent and old model references per directory.
Names the deprecated SDKs, google-generativeai and @google/generative-ai, so the agent installs google-genai or @google/genai instead.

Limitations

Needs a Gemini API key, and the managed-agent and Deep Research examples start remote sandboxes and long background jobs on the user's account.
Swaps in gemini-3.8-flash when a user asks for a deprecated model, and SKILL.md calls every gemini-2.5 model deprecated while the migration reference lists them as active legacy.
Much of the behavior lives in hosted docs fetched at run time, and the model list, agent IDs and minimum SDK versions are dated values that go stale quickly.

When not to use it

Real-time bidirectional audio or video, which the skill hands to the separate gemini-live-api-dev skill.
Languages other than Python and TypeScript, the only two its examples cover.

Reviewed by Retenta editorial on 2026-10-11; scope: source only.

Files included in this review

skills/gemini-api-dev/SKILL.md

Evidence and boundaries

Execution · not tested

Retenta has not executed this skill.

Security · not tested

The skill folder bundles no executable scripts. Retenta has not run a security scan.

Pinned publisher source: skills/gemini-api-dev/SKILL.md
Revision: 832c8f94114dc343d86ff27296fe9335b45affa0
License evidence: Apache-2.0Retenta preserves the license text captured with this immutable upstream snapshot; no independent legal or security review is claimed.
Last source change: 2026-10-06
Independent verification: Published by Google (Gemini team), but not independently verified by a third-party reviewer.
Upstream repository: Open publisher repository