Second Opinion

Get input from Gemini on the current problem or question. By default, iterate if the response lacks confidence.

Usage

/second-opinion <question or context>
/second-opinion --quick <question>        # Single pass, no iteration
/second-opinion --words=300 <question>    # Limit response to 300 words
/second-opinion --timeout=120 <question>  # Set timeout to 120s
/second-opinion                           # Uses current conversation context

Parameters

Parameter	Default	Description
--quick	off	Single pass, no iteration
--timeout	300	Timeout per advisor in seconds
--words	500	Max words per advisor response

How It Works

Default Flow (Iterative)

1. Summarize the current problem/question from the conversation, or use what the user provides
2. Query Gemini for its perspective
3. Evaluate confidence in the response
4. If confidence is LOW, re-query with additional context, up to 2 iterations
5. Present final results with your synthesis

Quick Mode (--quick)

1. Query Gemini once
2. Present results immediately without iteration
3. Use this when fast input matters more than refinement

Execution

Step 1: Prepare the Context

Extract or use the user's question/problem. If not explicitly provided, summarize:

• What is the current task or problem?
• What approaches are being considered?
• Any relevant file paths or code context

Write the prompt to .second-opinion.md in the current working directory. Use this exact filename for all subsequent steps:

Read-only consultation. Do not modify any files.

I need a second opinion: {problem_summary}

Give your perspective in {words} words or less. Focus on:
- Key considerations I might be missing
- Potential issues with the current approach
- Alternative approaches worth considering

If you need more context to give a confident answer, say so clearly.

Step 2: Query Gemini

Run Gemini using the Bash tool timeout set to {timeout} seconds:

gemini -p "Read the file at .second-opinion.md and follow the instructions within it." --approval-mode default --output-format text

If gemini is unavailable in the current environment, report that clearly rather than failing silently.

If Gemini prints status lines such as Loaded cached credentials. before the actual response, ignore those lines and use only the substantive model output.

Step 3: Evaluate Confidence

After receiving responses, evaluate each for confidence level:

High Confidence Indicators:

• Direct, specific recommendations
• References to specific code, files, or patterns
• Clear reasoning with concrete examples
• Definitive statements about approach

Low Confidence Indicators:

• Hedging language such as "it depends", "possibly", "might", or "could be"
• Requests for more context
• Generic advice that could apply to any situation
• Uncertainty markers such as "I'm not sure" or "hard to say"
• Questions back to you about the problem

Step 4: Iterate If Needed (Default Mode Only)

If confidence is LOW:

1. Identify what context is missing based on their feedback
2. Gather additional context by reading relevant files or clarifying requirements from the conversation
3. Overwrite .second-opinion.md with enhanced context
4. Re-query Gemini using the same file-reference command
5. Iterate at most 2 times

Skip this step entirely if --quick is used.

Step 5: Present Results

Format the responses for the user:

## Second Opinions

### Gemini
{gemini_response}

### My Take
{your brief synthesis - key takeaways and your recommendation}

If iteration occurred, note it:

*Note: Re-queried {advisor} with additional context after the initial response lacked confidence.*

Step 6: Clean Up

Delete .second-opinion.md when finished:

rm .second-opinion.md

Error Handling

• If Gemini fails, inform the user and offer to retry
• If the environment blocks gemini, surface that explicitly instead of silently omitting it

Examples

/second-opinion Should I use useCallback here or is it premature optimization?
/second-opinion Is this the right place to add error handling?
/second-opinion Review my approach to implementing this feature
/second-opinion --quick Just tell me if this pattern looks right
/second-opinion

Troubleshooting

Advisor times out or fails to respond

Solution: Increase the timeout with --timeout=600 or use --quick to skip iteration. If one advisor consistently fails, present the other advisor's response and mention the gap.

All advisors agree with no diversity of opinion

Solution: Rephrase the question to include a specific alternative you want evaluated, or provide more context about the trade-offs you are weighing.