Show HN: Issue tracking built for coding agents

3 weeks ago 1

Blazing-fast issue tracker for AI agents

A lightweight, dependency-aware issue tracker designed for AI coding agents. Track work, manage dependencies, and discover ready tasks—all from the CLI.

Imagine you're working on a project with multiple tasks, some blocking others. Tracer helps you:

📝 Track all your tasks in one place (no more scattered TODOs)
🔗 Link dependencies - "Task B can't start until Task A is done"
✅ See what's ready - Instantly find work you can start right now
🤖 Perfect for AI agents - They can track their own work across sessions
📦 Git-friendly - Everything stored as simple JSON, syncs via git

Real Example: You're building a login system. You need to design the database schema first, then implement the API, then build the UI. Tracer tracks these dependencies and always shows you what's actually ready to work on.

cargo install --git https://github.com/Abil-Shrestha/tracer

Prerequisites: Rust toolchain (installs in ~30 seconds)

💡 Tip: Use tr as a shorthand - both tracer and tr work identically!

# Initialize in your project tracer init # Learn how to use it tracer learn # Interactive tutorial showing all features

For AI Agents: See AGENTS.md for integration guide and best practices.

For Humans: Check out the Full documentation for detailed examples.

🎯 Example Workflow (with Dependencies)

Once you're comfortable with basics, try managing dependencies:

# Create an epic (big feature) tracer create "User authentication system" -t epic # Create subtasks tracer create "Design database schema" -t task tracer create "Build login API" -t task tracer create "Create login UI" -t task # Link them: API depends on schema being done tracer dep add test-3 test-2 --type blocks # See what you can work on RIGHT NOW tracer ready # → Shows test-2 (Design database schema) # → Hides test-3 (blocked by test-2) # Start working on what's ready tracer update test-2 --status in_progress

The power: Tracer automatically figures out what's ready to work on based on your dependencies!

⚡ Fast - ~5ms per operation, built in Rust
🤖 AI-friendly - JSON output, CLI-first, programmatic workflows
🔗 Smart dependencies - Track what blocks what, discover ready work
📦 Git-native - JSONL storage, no server needed
🌍 Distributed - Share work across agents via git
💾 Audit trail - Every change logged with context

Dependency Tracking

Four types: blocks, parent-child, discovered-from, related
Automatic ready work detection (finds unblocked tasks)
Dependency trees and cycle detection

AI Agent Integration

--json flag on all commands for programmatic parsing
Auto-export to JSONL after changes
Auto-import after git pull (hash-based)

Git-Based Storage

One JSON line per issue = clean git diffs
Commit .trace/issues.jsonl to version control
Clone repo = clone full database

Multi-Project Support

Auto-discovers database (like git)
Works from any subdirectory
Project-local isolation

# Core operations tracer create "Title" [-p priority] [-t type] tracer list [--status open] [--priority 1] tracer show <id> tracer update <id> --status in_progress tracer close <id> --reason "Done" # Dependencies tracer dep add <from> <to> --type blocks tracer dep tree <id> # Find work tracer ready [--limit 5] tracer blocked # Data management tracer export [-o file.jsonl] tracer stats

💡 Tip: Add --json to any command for programmatic parsing. Use tr instead of tracer for faster typing!

Fast enough to never slow you down:

Operation Time

Create issue	~5ms
List 1000 issues	~15ms
Ready work query	~10ms

Benchmarks on M1 MacBook Pro with 10,000 issues

Installation (Alternative Methods)

# From source git clone https://github.com/Abil-Shrestha/tracer.git cd tracer cargo install --path . # Verify installation tracer --version tr --version # Both commands work! # Using cargo (once published) cargo install tracer