Skills behavioral-modes
🧠

behavioral-modes

Safe

Adapt AI Behavior with Seven Operational Modes

Generic AI responses waste time and miss task-specific needs. This skill provides seven behavioral modes that optimize AI output for brainstorming, implementation, debugging, review, teaching, shipping, and multi-agent collaboration.

Supports: Claude Codex Code(CC)
📊 71 Adequate
1

Download the skill ZIP

2

Upload in Claude

Go to Settings → Capabilities → Skills → Upload skill

3

Toggle on and start using

Test it

Using "behavioral-modes". What if we added real-time collaboration to our document editor?

Expected outcome:

Let's explore this together. Here are some approaches:

Option A: WebSocket-based sync with operational transforms
Pros: Low latency, industry standard
Cons: Complex conflict resolution

Option B: CRDT-based architecture
Pros: Mathematically guaranteed consistency
Cons: Higher memory usage

Option C: Server-authoritative model
Pros: Simple implementation, clear ownership
Cons: Single point of failure

What resonates with your requirements?

Using "behavioral-modes". Review my authentication middleware

Expected outcome:

Code Review: auth.middleware.ts

Critical:
- JWT secret loaded from hardcoded string - use environment variable

Improvements:
- Add rate limiting to prevent brute force attacks
- Include token expiration validation

Good:
- Proper password hashing with bcrypt
- Input validation on credentials

Security Audit

Safe
v1 • 2/25/2026

All 34 static analysis findings are false positives. The SKILL.md file is pure documentation describing behavioral modes with markdown-formatted output examples. No executable code, shell commands, or cryptographic operations exist. The skill only uses read-only tools (Read, Glob, Grep) and provides behavioral guidance for AI interactions.

1
Files scanned
248
Lines analyzed
0
findings
1
Total audits
No security issues found
Audited by: claude

Quality Score

38
Architecture
100
Maintainability
87
Content
31
Community
100
Security
91
Spec Compliance

What You Can Build

Development Team Lead

Use DEBUG mode for systematic bug investigation and REVIEW mode for thorough pull request analysis with severity-categorized feedback.

Product Designer

Use BRAINSTORM mode to explore multiple feature alternatives with visual diagrams before committing to implementation direction.

Learning Developer

Use TEACH mode for concept explanations with analogies and exercises, then IMPLEMENT mode for production-ready code examples.

Try These Prompts

Basic Mode Activation
Use BRAINSTORM mode to generate three different approaches for implementing user authentication in my React app.
Debug Workflow
My login API returns 401 errors even with correct credentials. Use DEBUG mode to investigate: check the auth middleware, trace the token validation flow, and identify the root cause.
Code Review Request
Review this pull request using REVIEW mode. Categorize findings by severity (Critical, High, Medium, Low) and provide improved code examples for each issue.
Multi-Agent PEC Pattern
Use the PLAN-EXECUTE-CRITIC pattern: First create a task breakdown in task.md, then implement the feature using IMPLEMENT mode, finally perform security and performance review using REVIEW mode.

Best Practices

  • Explicitly request a mode when the task type is clear to ensure optimal AI behavior
  • Combine BRAINSTORM mode early with IMPLEMENT mode later for structured workflow
  • Use SHIP mode checklist before any production deployment

Avoid

  • Using IMPLEMENT mode for open-ended exploration that needs BRAINSTORM mode
  • Skipping REVIEW mode when deploying security-sensitive features
  • Expecting mode switching to bypass AI safety guidelines or tool limitations

Frequently Asked Questions

How does automatic mode detection work?
The skill matches keywords in your request to trigger words. For example, 'ideas' or 'what if' triggers BRAINSTORM mode, while 'error' or 'not working' triggers DEBUG mode. You can always explicitly request a mode using /mode syntax.
Can I combine multiple modes in one session?
Yes. A common pattern is BRAINSTORM for ideation, then IMPLEMENT for coding, then REVIEW for quality checks. Each request can use a different mode based on your current task phase.
What happens if the wrong mode is detected?
Simply restate your request with an explicit mode prefix like '/debug' or 'Use DEBUG mode'. The explicit request overrides automatic detection.
Do modes affect which tools the AI can use?
No. Modes only change communication style, output format, and prioritization. Tool availability depends on the skill's allowed-tools configuration, not the active mode.
Is TEACH mode suitable for production code generation?
TEACH mode prioritizes educational clarity over production optimization. Use IMPLEMENT mode for production code, then TEACH mode separately if you need explanation of the generated code.
How does the multi-agent PEC pattern work?
PEC (Plan-Execute-Critic) sequences three modes: Planner decomposes tasks into steps, Executor implements using IMPLEMENT mode, and Critic reviews using REVIEW mode. This pattern improves quality on complex tasks.

Developer Details

File structure

📄 SKILL.md