Analyze impact

Analyze Impact Command

Analyze the blast radius and dependencies for a specific file, feature, or set of changes. Helps understand what might break before making changes.

Arguments

$ARGUMENTS

If no argument provided, analyze currently staged or uncommitted changes.

Implementation

Step 1: Identify Target

Determine what to analyze:

• If a file path or directory path is provided: Analyze those files.
• If a free-text name is provided (e.g. a feature name or component name): use git ls-files and grep to find files matching that name in their path or contents, then analyze the most relevant set.
• If no argument: Analyze git diff --name-only (uncommitted changes).

Step 2: Parallel Dependency Analysis

Use the Task tool to launch multiple Explore agents simultaneously:

# Send a SINGLE message with THREE parallel Task calls:

Task 1: {
  subagent_type: "Explore",
  prompt: "Find all files that IMPORT or DEPEND ON [target files].
           Trace the dependency graph outward. List all affected files."
}

Task 2: {
  subagent_type: "Explore",
  prompt: "Find all files that [target files] IMPORT or DEPEND ON.
           Trace the dependency graph inward. Identify core dependencies."
}

Task 3: {
  subagent_type: "Explore",
  prompt: "Find all TESTS that cover [target files].
           Identify test files, test functions, and coverage areas."
}

Collecting the results. Follow ${CLAUDE_PLUGIN_ROOT}/skills/review/fan-in-protocol.md: stamp a deadline at dispatch, and when it expires, produce the analysis with whatever reported rather than waiting indefinitely on a silent agent.

An agent that never reported is not an empty result — and here the difference is dangerous. If the outward-dependency agent goes silent and you render its silence as "no dependents found", the report says this change is safe to make precisely when you have the least evidence for that. Blast radius is the one output where a gap and a clean result look identical and mean opposite things. Label any missing direction explicitly — dependents: unknown — agent did not report — and state in the summary that the impact assessment is partial.

Step 3: Expert Analysis (if MCP tools available)

If mcp__pal__analyze is available:

mcp__pal__analyze with:
  analysis_type: "architecture"
  relevant_files: [target files and dependencies]
  step: "Analyze impact and blast radius of changes to these files"

Step 4: Generate Impact Report

Impact Report Format

# Impact Analysis Report
**Target**: [file/feature being analyzed]
**Generated**: [timestamp]

## Summary
- **Direct Dependencies**: X files depend on this
- **Indirect Dependencies**: Y files affected transitively
- **Test Coverage**: Z test files cover this code
- **Risk Level**: [Low/Medium/High/Critical]

## Dependency Graph

### Files That Import This (Dependents)
Files that will break if this file's exports change:

| File | Import Type | Risk |
|------|-------------|------|
| `src/components/UserProfile.tsx` | Named import: `useAuth` | High |
| `src/pages/Login.tsx` | Default import | Medium |

### Files This Imports (Dependencies)
Files that this code relies on:

| File | Import Type | Stability |
|------|-------------|-----------|
| `src/utils/api.ts` | Named import | Stable |
| `src/types/user.ts` | Type import | Stable |

## Test Coverage

### Direct Tests
Tests that specifically test this file:
- `tests/auth/useAuth.test.ts` - 15 test cases
- `tests/integration/auth.test.ts` - 8 test cases

### Indirect Tests
Tests that exercise this code through other paths:
- `tests/e2e/login.test.ts` - Uses auth flow

### Coverage Gaps
Areas with insufficient test coverage:
- Error handling in `refreshToken()`
- Edge case: expired session

## Breaking Change Analysis

### Public API Surface
Exports that other files depend on:
```typescript
export function useAuth(): AuthContext  // 12 dependents
export type AuthState = {...}           // 8 dependents
export const AUTH_EVENTS = {...}        // 3 dependents

Safe Changes

These modifications are unlikely to break dependents:

• Internal implementation changes
• Adding new exports
• Performance optimizations

Risky Changes

These modifications may break dependents:

• Changing function signatures
• Modifying return types
• Removing or renaming exports

Recommendations

Before Making Changes:

1. Run affected tests: npm test -- --grep "auth"
2. Review all dependent files listed above
3. Consider feature flag for breaking changes
4. Update types if changing API surface

After Making Changes:

1. Run full test suite
2. Check for TypeScript errors in dependents
3. Update documentation if API changed
4. Consider deprecation warnings for removed APIs

Related Knowledge Files (if /ccmagic:map-codebase has been run)

• Architecture: context/knowledge/ARCHITECTURE.md
• Conventions: context/knowledge/CONVENTIONS.md
• Stack: context/knowledge/STACK.md

## Quick Analysis Mode

For quick checks, run a simplified analysis:

```bash
# Find direct imports (grep-based, fast)
grep -r "from.*[target-file]" --include="*.ts" --include="*.tsx" --include="*.js"

# Find test files
find . -name "*.test.*" -o -name "*.spec.*" | xargs grep -l "[target-name]"

Integration Points

• Before refactoring: Understand what might break
• Before deleting code: Ensure nothing depends on it
• Planning features: Understand integration complexity
• Code review: Assess change risk

Usage Examples

# Analyze a specific file
/ccmagic:analyze-impact src/auth/useAuth.ts

# Analyze a feature
/ccmagic:analyze-impact authentication

# Analyze current changes
/ccmagic:analyze-impact

Execution

Begin analysis immediately. Use parallel Explore agents for efficiency. Focus on actionable insights that help developers understand change risk before modifying code.