Skip to main content
Prompts Code Second Opinion from Gemini and Codex

agent coding workflow risk: medium

Code Second Opinion from Gemini and Codex

When invoked, summarizes the coding problem from conversation context in ~100 words, spawns gemini-consultant and codex-consultant subagents in parallel using the summary, and pres…

  • Policy sensitive
  • Human review
  • External action: medium

PROMPT 

---
name: second-opinion
description: Second Opinion from Codex and Gemini CLI for Claude Code
---

# Second Opinion

When invoked:

1. **Summarize the problem** from conversation context (~100 words)

2. **Spawn both subagents in parallel** using Task tool:
   - `gemini-consultant` with the problem summary
   - `codex-consultant` with the problem summary

3. **Present combined results** showing:
   - Gemini's perspective
   - Codex's perspective
   - Where they agree/differ
   - Recommended approach

## CLI Commands Used by Subagents

```bash
gemini -p "I'm working on a coding problem... [problem]"
codex exec "I'm working on a coding problem... [problem]"
```

REQUIRED CONTEXT

  • conversation context

TOOLS REQUIRED

  • task

ROLES & RULES

  1. Summarize the problem from conversation context (~100 words)
  2. Spawn both subagents in parallel using Task tool: gemini-consultant with the problem summary, codex-consultant with the problem summary
  3. Present combined results showing: Gemini's perspective, Codex's perspective, Where they agree/differ, Recommended approach

EXPECTED OUTPUT

Format
structured_report
Schema
numbered_steps · Summarize the problem, Spawn both subagents, Present combined results
Constraints
  • Gemini's perspective
  • Codex's perspective
  • where they agree/differ
  • recommended approach

SUCCESS CRITERIA

  • Summarize the problem accurately from context
  • Spawn and consult both subagents
  • Present balanced comparison of perspectives
  • Recommend an approach based on opinions

FAILURE MODES

  • Incomplete or inaccurate problem summary
  • Failure to spawn subagents in parallel
  • Unbalanced presentation of perspectives
  • Lack of clear agreements/differences

CAVEATS

Dependencies
  • conversation context
  • Task tool
  • gemini-consultant subagent
  • codex-consultant subagent
  • CLI commands for subagents
Missing context
  • Definition or usage instructions for 'Task tool'.
  • Prerequisites for subagents 'gemini-consultant' and 'codex-consultant'.
  • Access method for 'conversation context'.
  • Handling for subagent failures or timeouts.
Ambiguities
  • Assumes 'Task tool' and subagents 'gemini-consultant'/'codex-consultant' are predefined without explanation.
  • 'Spawn both subagents in parallel' lacks exact syntax or method.
  • CLI commands are labeled 'Used by Subagents' but unclear if main agent or subagents execute them.

QUALITY

OVERALL
0.78
CLARITY
0.85
SPECIFICITY
0.75
REUSABILITY
0.80
COMPLETENESS
0.70

IMPROVEMENT SUGGESTIONS

  • Add explicit instructions or examples for using the Task tool to spawn subagents.
  • Define or link to subagent prompts/configurations.
  • Provide a structured output format template for the combined results.
  • Include error handling steps, e.g., 'If a subagent fails, note it and proceed with available responses.'

USAGE

Copy the prompt above and paste it into your AI of choice — Claude, ChatGPT, Gemini, or anywhere else you're working. Replace any placeholder sections with your own context, then ask for the output.

MORE FOR AGENT