grits-issue-tracker

Grits Issue Tracker Skill

Goal

To efficiently manage project tasks and utilize topological code context for informed development, leveraging Grits' Git-native and local-first architecture with AI agent automation features.

Instructions

1. Project State & Session Hydration

• Quick Status: gr pulse - Shows current work, blockers, and recent commits in clean JSON
• Context-Aware: Most commands automatically use the focused issue when no ID is specified

2. Issue Management (Automated)

• List Issues:
  • gr list - Human-readable table
  • gr list --format ids - Just IDs (one per line, perfect for scripting)
  • gr list --format json - Structured JSON for parsing
• Create with Auto-Discovery:
  • gr create "Fix auth bug" --scan-file "src/auth.rs" --scan-file "src/user.rs" --start-work
  • Automatically discovers symbols, adds them to issue, and starts work
• Update with File Scanning:
  • gr update --scan-file "path/to/file.rs" - Auto-discovers and adds symbols from files
  • Uses focused issue automatically (no --id needed after gr workon)

3. Context-Aware Code Analysis

• Smart Star Neighborhood:
  • gr star - Uses focused issue's symbols automatically (no file argument needed)
  • gr star --format symbols - One symbol per line (cleanest for agents)
  • gr star --format files - One file per line
  • gr star --format structured - Categorized JSON (default)
• Enhanced Context Assembly:
  • gr context assemble - Uses focused issue automatically
  • gr context assemble --auto-expand - Includes star neighborhoods of all symbols
  • gr context assemble --auto-expand --format json - Machine-readable format

4. Workflow Automation

• One-Shot Issue Creation: gr create "title" --scan-file "file1.rs" --scan-file "file2.rs" --start-work
• Sticky Focus: After gr workon <id>, all commands auto-target that issue
• Clean Outputs: All commands provide agent-friendly structured output with noise filtering

5. Setup & Maintenance

• Initialization: gr onboard prepares the local environment
• Analytics: gr stats provides high-level overview
• Configuration: gr config [set|get|list] manages settings
• Portability: gr export and gr import manage JSONL source of truth

Agent Workflow Examples

Efficient Issue Creation and Context Assembly:

gr create "Fix auth bug" --scan-file "src/auth.rs" --scan-file "src/user.rs" --start-work
gr star --format symbols  # Uses focused issue automatically
gr context assemble --auto-expand  # Uses focused issue + expands context

This workflow creates an issue, discovers symbols, starts work, and provides full context in just 3 commands.

Output Format Examples

Clean Star Output:

{
  "connected_files": ["src/auth.rs", "src/user.rs"],
  "connected_symbols": [
    "src/auth.rs::validate_user",
    "src/user.rs::User",
    "src/auth.rs::AuthError"
  ],
  "depth": 1
}

Simple Symbol List:

src/auth.rs::validate_user
src/user.rs::User
src/auth.rs::AuthError

Clean Pulse Output:

{
  "status": "ready",
  "current_work": {
    "id": "gr-abc123",
    "title": "Fix auth bug", 
    "status": "in-progress",
    "connected_files": ["src/auth.rs", "src/user.rs"]
  },
  "recent_commits": ["a1b2c3d Fix login validation"]
}

Learning by Example

Review the files in the examples/ directory for authentic input/output signatures:

• JSON Pulse: examples/pulse_example.txt
• Issue & Symbol Linkage: examples/workon_example.txt
• Graph Results: examples/star_example.txt
• Utilities & Setup: examples/utility_example.txt

Constraints

• Path Consistency: Always use forward slashes (/) for all file paths in gr commands
• Tool Usage: Use only the gr CLI via run_command
• Data Integrity: Do not modify .grits/ directly
• Context-Aware: Leverage sticky focus - most commands work without explicit IDs after gr workon