/discover

command: discover
disable-model-invocation: true
description: "Discovery phase - Multi-AI research and exploration"
aliases:
  - probe
  - research-phase

Discover - Discovery Phase 🔍

🤖 INSTRUCTIONS FOR CLAUDE

MANDATORY COMPLIANCE — DO NOT SKIP

**When the user explicitly invokes `/octo:discover`, you MUST execute the structured workflow below.** You are PROHIBITED from doing the task directly, skipping the multi-provider research phase, or deciding the task is "too simple" for this workflow. The user chose this command deliberately — respect that choice.

EXECUTION MECHANISM — NON-NEGOTIABLE

**You MUST execute this command by running the discovery workflow through `orchestrate.sh`. You are PROHIBITED from:**

• ❌ Using the Agent tool to research/implement yourself instead of invoking the workflow
• ❌ Using WebFetch/Read/Grep as a substitute for multi-provider dispatch
• ❌ Skipping `orchestrate.sh` calls because "I can do this faster directly"
• ❌ Implementing the task using only Claude-native tools (Agent, Write, Edit)

**Multi-LLM orchestration is the purpose of this command.** If you execute using only Claude, you've violated the command's contract.

---

When the user invokes this command (e.g., `/octo:discover <arguments>`):

**✓ CORRECT - Run the orchestrated probe workflow:**

bash "${CLAUDE_PLUGIN_ROOT:-${CODEX_PLUGIN_ROOT:-${HOME}/.claude-octopus/plugin}}/scripts/orchestrate.sh" probe <user's arguments>

Run this from the user's project directory — never `cd` into the plugin first. Dispatched providers sandbox file access to the invoking directory, so a plugin cwd makes the user's project files unreadable to every provider.

**✗ INCORRECT:**

Skill(skill: "octo:discover", ...)  ❌ Wrong here; recursive Skill dispatch can reload this command
Task(subagent_type: "octo:discover", ...)  ❌ Wrong! This is a command workflow, not an agent type

Step 1: Ask Clarifying Questions

**CRITICAL: Before starting discovery, use the AskUserQuestion tool to gather context:**

Ask 3 clarifying questions to ensure focused research:

AskUserQuestion({
  questions: [
    {
      question: "How thorough should the research be?",
      header: "Research Intensity",
      multiSelect: false,
      options: [
        {label: "Quick (1-2 min)", description: "2 agents — fast problem space scan"},
        {label: "Standard (2-4 min)", description: "4-5 agents — balanced multi-perspective coverage (recommended)"},
        {label: "Deep (3-6 min)", description: "6-7 agents — exhaustive analysis with web search"}
      ]
    },
    {
      question: "What's your primary focus area?",
      header: "Focus",
      multiSelect: false,
      options: [
        {label: "Technical implementation (Recommended)", description: "Code patterns, APIs, architecture"},
        {label: "Best practices", description: "Industry standards and conventions"},
        {label: "Ecosystem & tools", description: "Libraries, frameworks, community"},
        {label: "Trade-offs & comparisons", description: "Pros/cons analysis"}
      ]
    },
    {
      question: "How should the output be formatted?",
      header: "Output",
      multiSelect: false,
      options: [
        {label: "Detailed report (Recommended)", description: "Comprehensive write-up"},
        {label: "Summary", description: "Concise key findings"},
        {label: "Comparison table", description: "Side-by-side format"},
        {label: "Recommendations", description: "Actionable next steps"}
      ]
    }
  ]
})

Map the intensity answer:

• "Quick" → `quick`
• "Standard" → `standard`
• "Deep" → `deep`

After receiving answers, incorporate them into the `orchestrate.sh probe` invocation.

Step 2: Run Discovery via orchestrate.sh

bash "${CLAUDE_PLUGIN_ROOT:-${CODEX_PLUGIN_ROOT:-${HOME}/.claude-octopus/plugin}}/scripts/orchestrate.sh" --intensity=<quick|standard|deep> probe <user's arguments>

Example: `bash "${CLAUDE_PLUGIN_ROOT:-${CODEX_PLUGIN_ROOT:-${HOME}/.claude-octopus/plugin}}/scripts/orchestrate.sh" --intensity=standard probe OAuth authentication patterns`

Pass the intensity only as the `--intensity=` flag. A bracket token such as `[intensity=deep]` is not an orchestrate.sh option: bash reads it as prompt text and runs at standard intensity, and zsh rejects it as an unmatched glob.

Step 3: Post-Completion — Interactive Next Steps

**CRITICAL: After discovery completes, you MUST ask the user what to do next. Do NOT end the session silently.**

AskUserQuestion({
  questions: [
    {
      question: "Discovery phase complete. What would you like to do next?",
      header: "Next Steps",
      multiSelect: false,
      options: [
        {label: "Move to Define phase", description: "Scope and clarify requirements based on findings (/octo:define)"},
        {label: "Go deeper on a specific finding", description: "Research a particular area in more detail"},
        {label: "Run the full workflow", description: "Continue through all remaining phases (/octo:embrace)"},
        {label: "Export the research", description: "Save findings as a document"},
        {label: "Done for now", description: "I have what I need"}
      ]
    }
  ]
})

---

**Runs the `probe` workflow for the research/discovery phase.**

Quick Usage

Just use natural language:

"Research OAuth authentication patterns"
"Explore caching strategies for high-traffic APIs"
"Investigate microservices best practices"

What Is Discover?

The **Discover** phase of the Double Diamond methodology:

• Divergent thinking
• Broad exploration
• Multi-perspective research
• Problem space understanding

What You Get

• Multi-AI research (Claude + Antigravity + Codex)
• Comprehensive analysis of options
• Trade-off evaluation
• Best practice identification
• Implementation considerations

When To Use

• Starting a new feature
• Researching technologies
• Exploring design patterns
• Unders