Agent skill
gemini-advisor
Gets Gemini's independent opinion on any file, topic, or discussion. Use when users say "ask gemini", "gemini opinion", or "what does gemini think". Provides a second perspective from a different AI model with Claude's commentary. Requires GEMINI_API_KEY environment variable to be set.
Install this agent skill to your Project
npx add-skill https://github.com/kwiggen/claude-code-plugin/tree/main/skills/gemini-advisor
SKILL.md
Gemini Advisor Skill
Before Starting
- Check Gemini availability — Run
test -n "$GEMINI_API_KEY"via Bash to verify the API key is configured. If it fails, inform the user that Gemini API key is not configured and offer to answer the question directly. - Determine context — Figure out what the user wants Gemini's opinion on:
- A specific file → read the file content
- A code question → formulate the question with relevant context
- A discussion/decision → summarize the current discussion state
- A general topic → pass the question directly
Process
Step 1: Prepare the Query
Build a clear, self-contained prompt for Gemini that includes:
- The user's question or topic
- Relevant context (file contents, discussion summary, code snippets)
- Framing: "Provide your independent technical opinion on the following."
Step 2: Send to Gemini
Use Bash to invoke the Gemini API entrypoint:
echo "<context>" | node {pluginDir}/dist/gemini/cli.js --prompt "<question with framing>"
Or without stdin if no file/code context is needed:
node {pluginDir}/dist/gemini/cli.js --prompt "<full question with context>"
The CLI outputs JSON with success, output, model, and optional error fields.
Step 3: Present Results
Output Format
Gemini's Take
[Gemini's full response, formatted cleanly]
Claude's Commentary
[Brief commentary on Gemini's response — agreements, disagreements, additional considerations, or alternative perspectives. Keep this concise — the point is to get a second opinion, not to debate.]
Fallback Behavior
If Gemini API key is not configured or the call fails:
- Inform the user: "Gemini API key not configured."
- Offer to answer the question directly as Claude.
- Do NOT block on Gemini unavailability.
Guidelines
- Present Gemini's output faithfully — do not rewrite or censor it.
- Claude's commentary should be brief and additive, not defensive.
- If the user's question is about a file, always read and pipe the file content.
- Keep Gemini prompts self-contained — Gemini has no conversation context.
- Do not send sensitive information (secrets, credentials) to Gemini.
When to Use This Skill
Use gemini-advisor for second opinions on any topic, file, or technical question. Use gemini-reviewer (/gemini-review) for structured dual code reviews with formal synthesis and verdict.
CLI Reference
If unsure about Gemini CLI flags, run:
node {pluginDir}/dist/gemini/cli.js --help
Recommended Agent Skills
Expand your agent's capabilities with these related and highly-rated skills.
gemini-reviewer
Dual code review: gets independent reviews from both Claude and Gemini API, then synthesizes agreements, unique findings, and a verdict. Use when users say "gemini review", "dual review", or want a second-perspective code review. Requires GEMINI_API_KEY environment variable to be set.
pr-writer
This skill should be used when the user asks to "create a PR", "open a pull request", "write a PR description", "push and create PR", or needs to generate a best-in-class pull request description based on commits. Provides industry-standard PR description formatting following conventional commits, Microsoft Engineering Playbook, and HackerOne best practices.
code-reviewer
Reviews code for logic flaws, edge cases, performance bottlenecks, security vulnerabilities, maintainability concerns, architecture best practices, and adherence to local patterns and CLAUDE.md guidelines. Use when users ask for code review or analysis, including reviewing uncommitted changes, the last local commit, unpushed code, or a specific GitHub pull request.
antipattern-detector
Detects common failure patterns in proposals, architectures, plans, and decisions. Use when reviewing any plan to identify red flags before they become problems. Covers architecture, timeline, team, process, and technology anti-patterns. Works on any domain — not just code.
github-insights
Provides team GitHub activity insights including merged PRs, contributor leaderboards, merge velocity, review participation, and PR size analysis. Use when users ask about team productivity, who merged PRs, code contribution stats, review patterns, or developer activity.
teach-me
Pedagogical code walkthrough that teaches developers about their own code. Explains choices, tradeoffs, alternatives, and refactoring opportunities. Use when users say "teach me", "explain my changes", "walk me through", "what did I build", or want to understand their code at a deeper level. Works on uncommitted changes, last commit, unpushed commits, or a PR.
Didn't find tool you were looking for?