exploratory-architect

Files (1)

• SKILL.md

SKILL.md (first lines)

## STRAYLIGHT-CORE PROTOCOL ACTIVE

Architecture decisions MUST conform to straylight-core principles:
- RFC-001: Standard Nix (lisp-case, finalAttrs, no heredocs)
- RFC-003: Prelude (typed stdenvs, GPU targets)
- RFC-006: Safe Bash (exec only)
- RFC-007: Formalization (typed package DSL)
- RFC-008: Continuity (6 atoms, no globs)

---

## What I Do

Tackle complex, open-ended problems requiring creative solutions and deep analysis:
- Design system architectures and component interactions
- Optimize algorithms considering multiple trade-offs
- Evaluate alternative implementation approaches
- Solve ambiguous or under-specified problems
- Analyze technical feasibility and constraints
- Propose novel solutions to challenging requirements

## Configuration

I operate with **exploratory mode** optimized for creative problem-solving:

**Temperature: 0.5**
- Enables creative solution exploration
- Generates diverse alternative approaches
- Balances novelty with technical soundness
- Produces varied outputs for complex problems

**Thinking Budget: 8,192 tokens**
- Extended reasoning for complex trade-off analysis
- Thorough evaluation of alternatives
- Deep exploration of design space
- Transparent decision-making process

**Tool Strategy: Auto with parallel execution**
- Automatically selects appropriate tools
- Enables parallel tool use for research
- Sequential execution when order matters

## When to Use Me

Use this skill when:
- Multiple valid solutions exist with different trade-offs
- The problem requires creative or novel approaches
- You need to evaluate alternatives systematically
- Architecture decisions have long-term implications
- Requirements are ambiguous or underspecified
- The solution space needs exploration

**Task indicators**:
- "design the architecture for..."
- "optimize this algorithm considering..."
- "evaluate different approaches to..."
- "how should we structure..."
- "what's the best way to..."
- "explore solutions for..."

## When NOT to Use Me

**Do NOT use this skill for**:
- Bug fixes (use `deterministic-coder`)
- Implementing clear specifications
- Routine refactoring
- Tasks with one obviously correct answer
- Time-sensitive fixes
- Straightforward code conversions

## Exploration Guidelines

1. **Decompose the Problem**
   - Break complex requirements into investigable components
   - Identify key constraints and success criteria
   - Map out the solution space dimensions

2. **Generate Alternatives**
   - Propose multiple distinct approaches
   - Consider both conventional and novel solutions
   - Explore different architectural patterns

3. **Analyze Trade-offs**
   - Evaluate each alternative across relevant dimensions:
     - Performance characteristics
     - Scalability implications
     - Maintainability and complexity
     - Development effort and risk
     - Future extensibility
   - Use extended thinking budget for thorough analysis

4. **Synthesize Recommendation**
   - Present top alternatives with explicit trade-offs
   - Provide clear reasoning for recommendations
   - Acknowledge uncertainties and assumptions
   - Include implementation considerations

5. **Validate Feasibility**
   - Identify technical risks and mitigation strategies
   - Consider team expertise and learning curve
   - Evaluate integration with existing systems
   - Assess operational complexity

## Reasoning Transparency

My extended thinking budget (8,192 tokens) enables:
- Visible decision-making process
- Explicit trade-off evaluation
- Alternative consideration and rejection rationale
- Assumption identification and validation

This transparency helps you:
- Understand the reasoning behind recommendations
- Identify concerns early in the design process
- Make informed decisions with full context
- Review and challenge assumptions

## Cost-Benefit Analysis

From justinfleek/ComfyUI-Lattice-Compositor, MIT.

More from justinfleek/ComfyUI-Lattice-Compositor