Skip to content
Back to skills

Investigation Analysis

ASecurity

Analyze feature requests, refactor plans, or technical decisions to determine investment value and provide recommendations. Use when user mentions investigating, analyzing, or asks "should we build this".

  • 666 stars
  • 0 votes
  • 0 copies
  • 2 views
  • Added February 9, 2026
developmentgorefactoringperformance

Security analysis

A100/100

Scanned February 12, 2026

npx -y skills add majiayu000/claude-skill-registry --skill investigation-analysis --agent claude-code

Installs into .claude/skills of the current project.

Are you the author of Investigation Analysis?

Add the live security badge to your README. It updates with every re-scan.

Security grade badge for Investigation Analysis
[![Security: A — Skills Directory](https://www.skillsdirectory.com/api/skills/majiayu000-investigation-analysis/badge)](https://www.skillsdirectory.com/skills/majiayu000-investigation-analysis)

More formats (shields.io, HTML) on the badges page. Keep it an A: scan every change in CI with Pro.

Download with Pro
SKILL.md
---
name: Investigation & Analysis
description: Analyze feature requests, refactor plans, or technical decisions to determine investment value and provide recommendations. Use when user mentions investigating, analyzing, or asks "should we build this".
allowed-tools: Read, Grep, Glob
---

# Investigation & Analysis Skill

## Purpose
Analyze feature requests or refactor plans to determine investment value and provide actionable recommendations.

## Auto-Activation Triggers
This skill activates when the user:
- Mentions investigating a feature or idea
- Asks "should we build this?"
- Requests feasibility analysis
- Wants to evaluate ROI or cost/benefit
- Asks "is this worth doing?"
- Mentions analyzing a refactor plan

## Analysis Process

### 1. Context Gathering
**Automatically check:**
- Review relevant existing code/architecture
- Search memory patterns for similar work
- Identify affected components and dependencies
- Review procedural memory for proven patterns

**Tools to use:**
- `Grep` to search codebase for related functionality
- `Glob` to find relevant files
- `Read` to examine current implementation

### 2. Investment Assessment

Evaluate across three dimensions:

#### Technical Factors
- **Implementation Complexity** (1-10 scale)
- **Risk Assessment** (breaking changes, dependencies)
- **Technical Debt Impact** (reduces/increases/neutral)
- **Performance Implications**

#### Business Factors
- **User Value Delivered**
- **Alignment with Project Goals**
- **Time to Implement**
- **Return on Investment**

#### Strategic Factors
- **Architecture Impact**
- **Reusability Potential**
- **Future Flexibility**

### 3. Recommendation Framework

**Question 1: Is it worth the investment?**
- **YES** - High value, reasonable cost, low risk
- **NO** - Low value, high cost, or high risk
- **CONDITIONAL** - Worth it if specific conditions met

**Question 2: What should we do with the request?**
- **KEEP AS-IS** - Plan is solid
- **MODIFY** - Suggest specific improvements
- **PIVOT** - Recommend alternative approach
- **DEFER** - Not now, revisit when [conditions]
- **REJECT** - Clear reasons why not

### 4. Response Format

```markdown
## Investigation: [Feature/Refactor Name]

### Investment Analysis

**Worth the Investment:** [YES/NO/CONDITIONAL]

[Executive summary in 2-3 sentences]

**Key Metrics:**
- Complexity: [X/10]
- Implementation Time: [estimate]
- Risk Level: [Low/Medium/High]
- Value Delivered: [Low/Medium/High]
- ROI: [High/Medium/Low]

### Recommendation: [KEEP/MODIFY/PIVOT/DEFER/REJECT]

[Detailed explanation with supporting evidence]

#### Technical Analysis
[Key technical findings from codebase investigation]

#### Business Justification
[Value proposition and alignment with goals]

#### Proposed Modifications (if MODIFY)
1. [Specific change with rationale]
2. [Specific change with rationale]

#### Alternative Approach (if PIVOT)
[Description of better approach]

### Implementation Considerations

**Prerequisites:**
- [Required before starting]

**Success Criteria:**
- [Measurable outcome]

**Potential Blockers:**
- [Risk] → Mitigation: [strategy]

### Evidence & References
- Code files examined: [file paths]
- Similar patterns in memory: [references]
```

## Best Practices

### 1. Be Evidence-Based
- Reference actual code files examined
- Cite similar attempts from memory
- Include metrics where available

### 2. Be Pragmatic
- Focus on practical impact
- Consider current capacity and priorities
- Balance ideal vs. practical constraints

### 3. Provide Actionable Guidance
- Specific next steps if proceeding
- Clear reasons if not proceeding
- Measurable success criteria

### 4. Check Memory First
Always consult:
- `.claude/memory/active/quick-reference.md`
- `.claude/memory/active/procedural-memory.md`

### 5. Leverage Existing Work
- Search for similar features already implemented
- Identify reusable patterns and components
- Check if request duplicates existing functionality

## Integration

**After Investigation:**
- If approved → Suggest `/orchestrate-tasks` for implementation
- If complex → Suggest `/plan-as-group` for collaborative planning

**Update Context:**
- Document investigation results
- Add insights to procedural memory if reusable

## Examples

### Feature Request
**User:** "Should we add real-time collaboration to the editor?"

**Skill:**
1. Searches codebase for existing editor architecture
2. Checks memory for similar implementations
3. Evaluates WebSocket/polling options
4. Assesses complexity vs. user value
5. Provides recommendation with implementation path

### Refactor Plan
**User:** "I'm thinking about refactoring to async/await"

**Skill:**
1. Examines current implementation
2. Identifies bottlenecks
3. Assesses migration complexity and risk
4. Evaluates performance benefits
5. Recommends phased approach or alternative

## Skill Metadata

**Version:** 1.0.0
**Category:** Planning & Decision Support

Attribution

Is this your skill, or is something wrong with this listing? Request removal or report an issue. Author removals are honored within 72 hours.

Comments

Loading comments…

Newsletter

Get the weekly list of safe new skills, and the ones we pulled.