cxgrd input

Analyze the blast radius of a planned change.

Usage

cxgrd input "description of the change"

Before touching any code, describe what you're planning to change. CXGRD will tell you every file that will be affected, why it's affected, and how severe the impact is.

Example output

PS C:\Users\name\project1> cxgrd input "removing oauth.py"

═════════════════════════
  Blast Radius Analysis
═════════════════════════

ℹ  Analyzing change: "removing oauth.py"
ℹ  Detected 2 changed file(s)


▶ Impact Summary
────────────────

   Overall Risk:  MEDIUM
   Risk Score: 28/100
   [██████░░░░░░░░░░░░░░] 28%


▶ Affected Files
────────────────

   Direct impact: 1 file(s)
   Transitive impact: 8 file(s)
   Total affected: 9 file(s)

   Top affected files:

    CRITICAL  app/github_client.py
      └─ Depends on auth.py (direct)
    HIGH  app/analyzer/static.py
      └─ Depends on github_client.py (transitive, depth: 2)
    HIGH  app/web.py
      └─ Depends on github_client.py (transitive, depth: 2)
    MEDIUM  app/reviewer.py
      └─ Depends on static.py (transitive, depth: 3)
    MEDIUM  app/webhook.py
      └─ Depends on reviewer.py (transitive, depth: 3)
    MEDIUM  app/main.py
      └─ Depends on web.py (transitive, depth: 3)
    MEDIUM  tests/test_webhook.py
      └─ Depends on main.py (transitive, depth: 2)
    LOW  tests/test_analyzer.py
      └─ Depends on static.py (transitive, depth: 3)
    LOW  tests/test_reviewer.py
      └─ Depends on reviewer.py (transitive, depth: 3)


▶ Change Classification
───────────────────────

   • unknown: unknown change detected (30% confident)
   • unknown: unknown change detected (30% confident)


▶ Recommendations
─────────────────

   ⚠️  1 critical file(s) impacted — review carefully
   ✓ Run tests for directly affected modules

✓ Blast radius analysis saved to history

Use this before making any significant refactor or before handing a task off to an AI assistant.

AI agents and tools work best when they are given clear and specific instructions instead of generalized or vague instructions. For example, giving a command like this one:

cxgrd input " refactor the function ResolveConflict in file index.js to properly handle edge cases "

has high chances of producing more accurate output than this one:

cxgrd input " fix error in index.js "