Gemini Prompting
Prompt recipes and antipatterns for delegating coding work to Gemini 2.5
Test report
- Verdict
- Works with setup
- Score
- Tested
- Aug 10, 2026
- Environment
- Claude Code 2.x (agent harness)
Asked for a gemini-2.5-flash prompt that reviews a git diff for security issues only, cold and then following the skill, and the skill changed five specific things: it moved the task after the diff, went from one negative constraint to four, swapped the prose 'reply in JSON' for an API response_schema it explicitly calls fragile to do in prose, pinned temperature 0.3 and a 2048 thinking budget from its own tables, and flagged that structured output breaks on 2.5 when tool calls are in history. That is a real improvement to the prompt, but with no gemini binary on PATH and no API key there was no way to run either prompt, so the downstream answer quality is unmeasured. The install itself is clean from a bare HOME. Two structural catches: the frontmatter carries user-invocable: false, so once installed standalone you cannot call it yourself and must delete that line first, and the 590-line body never links its own references/ directory, so those two files sit on disk and never load. Content is still keyed to Gemini 2.5, with Gemini 3 surfacing only as footnotes that invert the skill's own temperature table.
Scored on four weighted criteria — install, triggering, output vs. baseline, docs. How scoring works
- Installs cleanly 5/5
- Triggers reliably 4/5
- Output vs. baseline 7/10
- Docs & honesty 3/5
What Gemini Prompting does
A single large reference on composing Gemini 2.5 Pro/Flash prompts for code review, bug diagnosis, refactors and codebase analysis — model choice, thinking budgets, temperature, structured output via response_schema, and ten named antipatterns. Loads when you are writing a prompt to hand a coding task to Gemini. Built as an internal component of the Gemini Claude Code plugin rather than a standalone skill.
How to install Gemini Prompting
git clone https://github.com/abiswas97/gemini-plugin-cc.git
mkdir -p ~/.claude/skills
cd gemini-plugin-cc && cp -r plugins/gemini/skills/gemini-prompting ~/.claude/skills/gemini-prompting
Skills live in ~/.claude/skills/ (global) or .claude/skills/
(per-project). Restart Claude Code after installing.
Commands — how to trigger Gemini Prompting
-
/gemini-promptingPrompt recipes and antipatterns for delegating coding work to Gemini 2.5
It also activates on plain-language prompts like these:
-
Write a Gemini 2.5 Flash prompt to review this diff for security -
Compose a Gemini prompt with response_schema for bug diagnosis -
Help me pick thinking budget and temperature for this Gemini task
Frequently asked questions
- Is the Gemini Prompting skill free?
- Yes. The skill itself is free from abiswas97/gemini-plugin-cc. SkillProof publishes the install command and an independent test verdict at no cost.
- Does Gemini Prompting work with Claude Code?
- We tested it with Claude Code 2.x (agent harness) on Aug 10, 2026. Verdict: Works with setup. Asked for a gemini-2.5-flash prompt that reviews a git diff for security issues only, cold and then following the skill, and the skill changed five specific things: it moved the task after the diff, went from one negative constraint to four, swapped the prose 'reply in JSON' for an API response_schema it explicitly calls fragile to do in prose, pinned temperature 0.3 and a 2048 thinking budget from its own tables, and flagged that structured output breaks on 2.5 when tool calls are in history. That is a real improvement to the prompt, but with no gemini binary on PATH and no API key there was no way to run either prompt, so the downstream answer quality is unmeasured. The install itself is clean from a bare HOME. Two structural catches: the frontmatter carries user-invocable: false, so once installed standalone you cannot call it yourself and must delete that line first, and the 590-line body never links its own references/ directory, so those two files sit on disk and never load. Content is still keyed to Gemini 2.5, with Gemini 3 surfacing only as footnotes that invert the skill's own temperature table.
- What is the Gemini Prompting SkillProof Score?
- 7.6/10 — installs cleanly 5/5, triggers reliably 4/5, output vs. baseline 7/10, docs & honesty 3/5.
- How do I install Gemini Prompting?
- Copy the install command from this page, run it in your terminal, and restart Claude Code. Skills live in ~/.claude/skills/ (global) or .claude/skills/ inside a project.
- Can I use Gemini Prompting with Cursor, Copilot, Gemini CLI, Codex or other AI tools?
- The SKILL.md format is native to Claude (Claude Code, Desktop, claude.ai). The instructions inside adapt to other assistants: Cursor rules, GitHub Copilot instructions, Windsurf rules, Custom GPTs, AGENTS.md for OpenAI Codex, and GEMINI.md for Google Gemini CLI — our conversion guides cover each, and the free converter on the tools page does the wrapping for you.