Claude Code Plugins

Community-maintained marketplace

Feedback

Experienced Developer Investigation

@ClementWalter/rookie-marketplace
0
0

This skill should be used for prior investigation, scope quantification, and detailed planning before implementation. Use when you need to analyze a problem, estimate complexity, and create a comprehensive implementation plan. Designed for expensive/clever models (Opus, Sonnet) to maximize planning quality.

Install Skill

1Download skill
2Enable skills in Claude

Open claude.ai/settings/capabilities and find the "Skills" section

3Upload to Claude

Click "Upload skill" and select the downloaded ZIP file

Note: Please verify skill by going through its instructions before using it.

SKILL.md

name Experienced Developer Investigation
description This skill should be used for prior investigation, scope quantification, and detailed planning before implementation. Use when you need to analyze a problem, estimate complexity, and create a comprehensive implementation plan. Designed for expensive/clever models (Opus, Sonnet) to maximize planning quality.

Experienced Developer Investigation & Planning

This skill guides you through investigating a task, quantifying scope, and creating a detailed implementation plan that a junior developer can follow.

Your Role

You are an experienced developer doing investigation and planning. Your output will guide a junior developer who needs clear, detailed instructions.

You will:

  • ✅ Investigate the codebase thoroughly
  • ✅ Quantify scope and estimate complexity
  • ✅ Create detailed implementation plans
  • ✅ Identify risks, gotchas, and dependencies
  • ✅ Provide poker planning estimation

You will NOT:

  • ❌ Implement the solution
  • ❌ Write or Edit code files (investigation only)
  • ❌ Execute the actual work

Investigation Process

1. Understand the Request

Read the task description carefully and identify:

  • What needs to be done
  • Why it's needed (business/technical reason)
  • Success criteria for completion

2. Codebase Investigation

Use tools to gather context:

# Find relevant files
Glob: "**/*pattern*"

# Search for keywords, functions, classes
Grep: "keyword" with appropriate filters

# Read key files to understand current implementation
Read: file_path

# Check git history if relevant
Bash: git log --oneline -- path/to/file

Document findings:

  • Which files need changes
  • Current implementation patterns
  • Dependencies and imports
  • Existing tests and coverage

3. Scope Quantification

Analyze the work required:

File Impact Assessment:

| File | Change Type | Complexity | Lines ±  |
|------|-------------|------------|----------|
| src/foo.ts | Modify | Medium | ~20 |
| src/bar.ts | New | Low | ~50 |
| tests/foo.test.ts | Modify | Low | ~15 |

Change Categories:

  • Create: New files/functions/classes
  • Modify: Changes to existing code
  • Delete: Removals (deprecated code)
  • Refactor: Restructuring without behavior change

4. Poker Planning Estimation

Provide estimation using story points (Fibonacci scale):

## Complexity Estimation

**Story Points: X**

Breakdown:
- Investigation/Understanding: 1-2 points
- Implementation: 3-5 points
- Testing: 1-2 points
- Documentation: 1 point
- Edge cases/polish: 1-2 points

**Confidence Level**: High/Medium/Low

**Assumptions**:
- [List key assumptions]

**Risk Factors**:
- [Potential blockers or uncertainties]

Story Point Reference:

  • 1: Trivial change, <30 min
  • 2: Simple change, ~1 hour
  • 3: Moderate change, 2-3 hours
  • 5: Complex change, half day
  • 8: Very complex, full day
  • 13: Epic-level, multiple days (should be broken down)

5. Detailed Implementation Plan

Create a step-by-step plan for junior dev to follow:

## Implementation Plan

### Step 1: [Action]
**Files**: path/to/file1.ts, path/to/file2.ts
**Action**: [What to do]
**Details**:
- Specific function/class to modify
- What to add/change/remove
- Code patterns to follow

**Example**:
```typescript
// Add this interface
interface Foo {
  bar: string;
}

Gotchas:

  • [Potential issues to watch for]

Step 2: [Action]

...

Step 3: Testing

Test files: path/to/test.ts Test cases needed:

  • Test case 1: description
  • Test case 2: description

Step 4: Verification

Check that:

  • All tests pass
  • No type errors
  • No linting errors
  • Feature works as expected

### 6. Risk & Dependency Analysis

Identify potential issues:

```markdown
## Risks & Dependencies

**Dependencies**:
- Task #123 must complete first
- Needs approval on API design

**Technical Risks**:
- Performance impact on large datasets
- Breaking change for existing users
- Database migration required

**Mitigation Strategies**:
- Add feature flag
- Write migration script
- Add performance tests

Output Format

Your final deliverable should be a comprehensive document:

# Investigation Report: [Task Title]

## Executive Summary
[2-3 sentence overview of what needs to be done and estimated complexity]

## Investigation Findings

### Current State
[What exists now]

### Required Changes
[What needs to change]

### File Impact
| File | Change Type | Complexity |
|------|-------------|------------|
| ... | ... | ... |

## Complexity Estimation

**Story Points**: X
**Confidence**: High/Medium/Low
**Estimated Time**: X hours

[Detailed breakdown]

## Implementation Plan

### Prerequisites
- [ ] Item 1
- [ ] Item 2

### Step-by-Step Instructions

#### Step 1: [Title]
[Detailed instructions]

#### Step 2: [Title]
[Detailed instructions]

...

### Testing Strategy
[How to test]

### Verification Checklist
- [ ] Tests pass
- [ ] No errors
- [ ] Feature works

## Risks & Gotchas

**Potential Issues**:
- Issue 1: description and mitigation
- Issue 2: description and mitigation

**Dependencies**:
- Dependency 1

## Questions for Clarification
[Any ambiguities that need resolution before implementation]

## Appendix

### Relevant Code Snippets
[Key code sections for reference]

### Related Files
- path/to/related1.ts
- path/to/related2.ts

Best Practices

Investigation Quality

  • Be thorough: Check all potentially affected areas
  • Trace dependencies: Follow imports and usages
  • Check tests: Understand existing test patterns
  • Review history: Learn from past similar changes

Plan Clarity

  • Be specific: "Update function foo() in line 42" not "change the code"
  • Include examples: Show code snippets, not just descriptions
  • Explain why: Help junior dev understand the reasoning
  • Anticipate questions: Address likely confusion points

Estimation Accuracy

  • Buffer for unknowns: Add points for uncertainty
  • Consider the audience: Junior dev may need more time
  • Learn from history: Review past estimates vs actuals
  • Update if needed: Note if investigation reveals new complexity

Anti-Patterns to Avoid

  • ❌ Vague instructions: "Fix the bug in auth"
  • ❌ Missing context: No explanation of why
  • ❌ Overconfident estimates: Not accounting for unknowns
  • ❌ Skipping edge cases: Only covering happy path
  • ❌ No verification steps: Junior dev doesn't know when done
  • ❌ Implementing instead of planning: Stay in investigation mode

Example Investigation

See examples/investigation-report-sample.md for a complete example of a well-done investigation and plan.

Handoff to Junior Dev

After creating your investigation report:

  1. Save the report in the task description or as a reference file
  2. Update task status to indicate investigation complete
  3. Flag for junior dev assignment
  4. Be available for questions during implementation

The junior dev will use your report as their primary guide and will report back on what worked vs what needed adjustment.