Show HN: Issue tracking built for coding agents (github.com)

Tracer is a lightweight, Rust-built CLI issue tracker purpose-built for AI coding agents and humans who want dependency-aware task management without a server. Announced on Show HN, it stores issues as one JSONL line per record in your repo, syncs via git, and exposes a programmatic interface (all commands support --json). That makes it easy for LLM-driven agents to track work across sessions, auto-import updates after a git pull, and discover ready tasks automatically based on explicit dependencies. Key technical highlights: Tracer is fast (~5–15ms ops in benchmarks with 10k issues), supports four dependency types (blocks, parent-child, discovered-from, related), does cycle detection, and computes “ready” work by resolving dependency trees. It logs an audit trail, is project-local (auto-discovers DB from any subdirectory), and provides CLI-first commands for create/list/show/update/dep/add/ready. Installation requires the Rust toolchain (cargo install from GitHub or local path). The agent-focused features—JSON output, auto-export/import, AGENTS.md integration guide, and git-native storage—make it a pragmatic choice for distributed multi-agent workflows and reproducible, auditable task orchestration in AI/ML engineering pipelines.