resolve-todos（解决待办事项）

Resolve TODOs Skill

Purpose: Resolve all TODO comments using parallel processing.

Arguments

• file-or-dir-path (optional): Path to a specific file or directory to scan for TODOs. If omitted, scans the entire repository.

Workflow

1. Gather TODOs

First, identify the scope:

• If an argument is provided, scan that specific file or directory
• If no argument, scan the entire codebase

Search for TODO comments in code files using Grep with pattern: (?i)TODO|FIXME|XXX

For each TODO found:

• Extract the TODO text
• Capture the file path and line number
• Include surrounding context (3-5 lines before and after) to understand the background

2. List TODOs

Present a comprehensive list of all TODOs with their context:

## Found N TODOs

### TODO 1: [Brief description]
**Location:** `path/to/file.ext:line`
**Context:**
\`\`\`language
// Surrounding code context
// With the TODO highlighted
\`\`\`

### TODO 2: ...

3. Prompt User Confirmation

Ask the user if they want to proceed with resolving all the TODOs:

"Found N TODOs. Do you want to proceed with resolving them? If yes, I'll use parallel processing to handle them efficiently."

Use AskUserQuestion to get confirmation before proceeding.

4. Plan Dependencies

Create a TodoWrite list of all TODOs. Analyze dependencies:

• Group TODOs by type (refactoring, feature addition, bug fix, etc.)
• Identify any dependencies between TODOs (e.g., one TODO requires another to complete first)
• Create a mermaid flow diagram showing the resolution order

Example mermaid diagram:

graph TD
    A[TODO: Rename variable X] --> B[TODO: Update references to X]
    C[TODO: Add new function] --> D[TODO: Write tests]
    B --> D

5. Resolve in Parallel

For each TODO that has no outstanding dependencies, spawn a general-purpose agent in parallel using the Task tool.

Example with 3 independent TODOs:

Launching parallel resolution agents for independent TODOs...

1. Spawning agent for TODO 1
2. Spawning agent for TODO 2
3. Spawning agent for TODO 3

Each Task should use subagent_type general-purpose with a prompt that includes:

• The TODO description
• The file location and context
• Clear instructions on what needs to be done

6. Commit & Resolve

After all parallel agents complete:

• Review the changes made
• Create a descriptive commit message summarizing all resolved TODOs
• Push to remote if configured

Notes

• The command searches for common TODO patterns: TODO, FIXME, XXX
• Context extraction helps understand the background and purpose of each TODO
• Parallel processing significantly speeds up resolution for independent TODOs
• Dependency analysis ensures correct resolution order