17th-code-analyzer

name: 17th-code-analyzer description: | Use this skill when user asks about code quality, duplications, dead code, or optimization opportunities. Analyzes staged changes AND surrounding code using GPT-5.3-Codex. Auto-triggers on: "중복 코드", "불필요한 코드", "최적화", "code quality", "duplications", "dead code" Appropriate for: code hygiene, pre-PR quality check, refactoring opportunities. Not appropriate for: running tests, writing code, code review for correctness.

Code Analyzer

Analyze staged code and surrounding context for duplications, dead code, and optimization opportunities using GPT-5.3-Codex.

Usage

/code-analyzer              → Codex high (default)
/code-analyzer deep         → Codex xhigh

Argument Parsing

Keywords (reserved):

• deep → Use xhigh effort instead of high

Parsing Logic

args = split(input)
effort = "high"

for arg in args:
    if arg == "deep":
        effort = "xhigh"

Result Mapping

Analysis Scope

Analysis Items

Implementation

Invoke the 17th-code-analyzer subagent:

Task(
    description="Analyze code quality ({effort})",
    subagent_type="17th-code-analyzer",
    prompt="""
Analyze code quality for staged changes and surrounding code.

Configuration:
- Effort: {effort}  (high or xhigh)
- Project root: {project_root}
""",
    run_in_background=true
)

Auto-Trigger

When skill auto-triggers (no args), uses defaults:

• Effort: high

Output

• Report: .kkachi/{branch_name}/completed/code-analyzer-{epoch_timestamp}.md
• Summary: duplications, dead code, optimizations found
• Recommendations: prioritized action items

Examples

Notes

• Uses GPT-5.3-Codex for all analysis
• Analyzes staged changes + surrounding code
• Supports Go, Python, Dart, TypeScript, JavaScript
• Results verified before reporting
• Ensure .kkachi/ is in .gitignore