Menu
Use for ideation, requirements clarification, and architectural design. Adapts to task complexity (Trivial/Medium/Complex).
Install with Codex or Claude Copy this prompt, paste it into Codex, Claude, or another assistant, and let it review the skill page and install it for you.
Based on SOC occupation classification
Use when decomposing tasks into parallel sub-tasks or spawning sub-agents. Vendor-agnostic core; load a per-vendor reference for concrete tool names, directory conventions, and invocation syntax.
Use when performing Verification-Driven Development with adversarial approach. Actively challenge assumptions and find weak spots.
Use when performing VDD adversarial review with an opt-in sarcastic, provocative delivery style — a stylistic skin over vdd-adversarial mechanics (exhaustive reporting + objective bar).
Performance critic in adversarial style (optional sarcastic skin). Part of VDD Multi-Adversarial pipeline.
Use when performing OWASP security critique in adversarial style (optional sarcastic skin). Part of VDD Multi-Adversarial pipeline.
Use when performing security vulnerability assessment (OWASP, secrets, dependencies, IaC, LLM, API, MCP/agentic) or when "thinking like a hacker" to find exploits.
| name | brainstorming |
| description | Use for ideation, requirements clarification, and architectural design. Adapts to task complexity (Trivial/Medium/Complex). |
| tier | 2 |
| version | 2.1 |
| status | active |
| changelog | Upgraded to Universal Gold Standard v2.1 (2026) - Added 3-Tier Assessment, Tool Agnosticism, and Safety Guardrails. |
Purpose: Bridge the gap between user intent ("Make it pop") and technical specification ("CSS Box Shadow"). Core Rule: Do NOT write code until you have a confirmed design (except for Trivial tasks).
1. Complexity Assessment (MANDATORY) At the start, classify the task to determine your mode:
| Level | Criteria | Protocol | Reference |
|---|---|---|---|
| TRIVIAL | Clear requirements, standard pattern, single component (e.g., "Fix typo", "Add log"). | Fast Path: Skip deep research. Confirm plan in chat. Go. | examples/demo_trivial.md |
| MEDIUM | Standard task but open details (e.g., "Add button", "Refactor function"). | Standard: 1-2 clarifying questions, 2 options, brief confirmation. | examples/demo_medium.md |
| COMPLEX | High uncertainty, architectural impact, multiple components (e.g., "New System"). | Deep Dive: Research -> Mermaid -> Trade-offs -> Design Doc -> Sign-off. | examples/demo_complex.md |
2. Context Gathering (Tool Agnostic)
grep, search_code) to understand existing patterns.
package.json and relevant files?"task.md, README.md, or .cursorrules to align with project standards.1. Smart Questions
2. Visual Thinking
3. The "Design Doc" Rule
docs/design/feature-name.md).
1. Adversarial CoT Checklist (The "Final Gate") Before asking for approval, explicitly check these in your internal thought process:
2. Handover Templates
| Agent Excuse | Reality / Correct Action |
|---|---|
| "It's faster to just code it." | WRONG. Reworking bugs is slower. Validate first. |
| "I skipped options because it's trivial." | OK, but state: "This is trivial, I propose X. Proceed?" |
| "The user ignored my question." | RISKY. Don't guess. Ask again: "To avoid breaking X, I need to know Y." |
| "I can't see the file." | STOP. Ask for it. Do not hallucinate file contents. |
| "This seems standard." | VERIFY. Is it standard for this project? Check package.json. |