/advise

name: advise
description: Pre-work command that queries past learnings and leverages MCPs before starting a new task
allowed-tools: Read, Glob, Grep, AskUserQuestion, mcp__context7__resolve-library-id, mcp__context7__get-library-docs, mcp__octocode__githubSearchCode, mcp__octocode__githubGetFileContent, mcp__octocode__githubViewRepoStructure, mcp__octocode__githubSearchRepositories

/own:advise

> ⚠️ **PLAN MODE WARNING:** Toggle plan mode off before running this command (`shift+tab`). OwnYourCode commands don't work correctly with plan mode.

Query past learnings and leverage research tools before starting a new task.

Overview

This command is run **before** starting new work. It: 1. Understands what you're about to work on 2. Queries your learning registry for relevant patterns/failures 3. Uses MCPs to research best practices and implementations 4. Surfaces past insights to prevent repeating mistakes 5. Ensures you have a plan before coding

> "Those who don't learn from history are doomed to rewrite it with bugs."

---

Execution Flow

Phase 1: What Are You About to Work On?

Question: "What are you about to work on?"

Options:
1. A new feature from my spec
   Description: Starting planned work

2. A bug fix
   Description: Something is broken

3. An improvement/refactor
   Description: Making existing code better

4. Something new (not specced)
   Description: Unplanned work

Follow-up:

• "In a few words, what's the domain? (e.g., auth, forms, API, database, UI)"
• "What libraries or frameworks are you using?"

---

Phase 2: Query Global Learning Registry

Read the **global** learning registry at `~/ownyourcode/learning/LEARNING_REGISTRY.md` and search for:

• Patterns matching the domain
• Failures matching the domain
• Recent learnings that might apply

Also check `~/ownyourcode/learning/patterns/` for any documented patterns in the domain.

**Important:** Learning is GLOBAL, not project-specific. This ensures learnings persist across all your projects.

---

Phase 3: Surface Relevant Learnings

If past learnings found:

┌─────────────────────────────────────────┐
│         RELEVANT PAST LEARNINGS         │
├─────────────────────────────────────────┤
│                                         │
│ 📚 PATTERNS YOU'VE USED                │
│ ─────────────────────────               │
│ • [Pattern Name] (2026-01-15)          │
│   "[Summary of what worked]"            │
│   Location: ~/ownyourcode/learning/...  │
│                                         │
│ ⚠️ FAILURES TO AVOID                   │
│ ─────────────────────────               │
│ • [Issue Name] (2026-01-01)            │
│   "[What went wrong and why]"           │
│   Location: ~/ownyourcode/learning/...  │
│                                         │
│ 💡 INSIGHTS                             │
│ ─────────────────────────               │
│ • "[Key insight from past work]"        │
│                                         │
└─────────────────────────────────────────┘

If no learnings found:

No specific learnings found for [domain] yet.

This will be one of your first deep dives into this area.
As you work, pay attention to:
- Patterns that work well
- Mistakes to avoid
- Insights worth remembering

Run /own:retro when you're done to capture learnings.

---

Phase 4: MCP-Powered Research

**Use ALL available MCPs to gather intelligence before coding:**

Context7 - Official Documentation

For any libraries/frameworks mentioned: 1. Resolve the library ID 2. Fetch latest documentation 3. Surface: Recent API changes, best practices, common pitfalls

> "Let me check Context7 for the latest [library] documentation..."

Octocode - GitHub Best Practices

Search GitHub for real-world implementations: 1. Search popular repositories for the pattern/feature 2. Find production-quality implementations 3. Surface: How top projects implement this, common patterns, tested approaches

Example queries:

• "How does [popular repo] implement [feature]?"
• "What's the standard pattern for [X] in [framework]?"
• "How do production apps handle [edge case]?"

> "Let me check Octocode for how production apps implement this..."

Use these Octocode tools:

• `githubSearchCode` - Find implementations across repos
• `githubGetFileContent` - Read specific implementations
• `githubSearchRepositories` - Find authoritative repos for reference

When to Use Each MCP

| Situation | Primary MCP | Secondary | |-----------|-------------|-----------| | Learning a library API | Context7 | Octocode (examples) | | Implementing a feature | Octocode | Context7 (docs) | | Best practices question | Octocode | - | | Debugging library issue | Context7 | Octocode (issues) | | Architecture decisions | Octocode | - |

---

Phase 5: Research Summary

Present findings in a structured way:

┌─────────────────────────────────────────┐
│           RESEARCH FINDINGS             │
├─────────────────────────────────────────┤
│                                         │
│ 📖 DOCUMENTATION (Context7)             │
│ ─────────────────────────               │
│ • [Library] v[X.Y.Z]                   │
│ • Key API: [relevant methods]           │
│ • Note: "[Important caveat]"            │
│                                         │
│ 🔍 PRODUCTION EXAMPLES (Octocode)       │
│ ─────────────────────────               │
│ • [Repo Name] implements this as:       │
│   "[Brief description of approach]"     │
│ • Common pattern: [Pattern observed]    │
│                                         │
│ ⚡ RECOMMENDATIONS                       │
│ ─────────────────────────               │
│ • Follow [repo]'s approach for [X]      │
│ • Avoid [anti-pattern seen in search]   │
│ • Consider [suggestion from research]   │
│                                         │
└─────────────────────────────────────────┘

---

Phase 6: Preparation Checklist

Before you start coding, confirm:

Question: "Before diving in, let's check p