2.8 KiB
2.8 KiB
| description | argument-hint |
|---|---|
| Critical thinking analysis - validates alignment, challenges assumptions, identifies risks | Proposal or decision to analyze (e.g., "Add Redis for caching") |
Critical Thinking Analysis
Invoke the advisor agent to analyze proposals, validate alignment, and identify risks before implementation.
Implementation
Invoke the advisor agent with the user's proposal.
ARGUMENTS: {{ARGUMENTS}}
Use the Task tool to invoke the advisor agent with subagent_type="advisor" and provide the proposal from ARGUMENTS.
What This Does
You describe a proposal or decision point. The advisor agent will:
- Validate alignment with PROJECT.md goals, scope, and constraints
- Analyze complexity cost vs benefit
- Identify technical and project risks
- Suggest simpler alternatives
- Provide clear recommendation (PROCEED/CAUTION/RECONSIDER/REJECT)
Time: 2-3 minutes
Usage
/advise Add Redis for caching
/advise Refactor to microservices architecture
/advise Switch from REST to GraphQL
/advise Add real-time collaboration features
Output
The advisor provides:
- Alignment Score (0-10): How well proposal serves PROJECT.md goals
- Decision: PROCEED / CAUTION / RECONSIDER / REJECT
- Complexity Assessment: Estimated LOC, files, time
- Pros/Cons: Trade-off analysis
- Alternatives: Simpler, more robust, or hybrid approaches
- Risk Assessment: What could go wrong
When to Use
Use /advise when making significant decisions:
- Adding new dependencies (Redis, Elasticsearch, etc.)
- Architecture changes (microservices, event-driven, etc.)
- Scope expansions (mobile support, multi-tenancy, etc.)
- Technology replacements (GraphQL vs REST, etc.)
- Scale changes (handling 100K users, etc.)
Integration
The advisor-triggers skill automatically suggests /advise when it detects significant decision patterns in your requests.
Next Steps
After receiving advice:
- PROCEED: Continue with
/planor/auto-implement - CAUTION: Address concerns, then proceed
- RECONSIDER: Evaluate alternatives before proceeding
- REJECT: Don't implement, or update PROJECT.md first
Comparison
| Command | Time | What It Does |
|---|---|---|
/advise |
2-3 min | Critical analysis (this command) |
/research |
2-5 min | Pattern and best practice research |
/plan |
3-5 min | Architecture planning |
/auto-implement |
20-30 min | Full pipeline |
Technical Details
This command invokes the advisor agent with:
- Model: Opus (deep reasoning for critical analysis)
- Tools: Read, Grep, Glob, Bash, WebSearch, WebFetch
- Permissions: Read-only analysis (cannot modify code)
Part of: Core workflow commands
Related: /plan, /auto-implement, advisor-triggers skill
GitHub Issue: #158