Coding · Reviewed 2026-07-23
graykode/abtop
STEADY · 82/100
A fast, actively-developed Rust 'htop for AI coding agents' — an excellent human-facing observability tool, not an agent-consumable component.
Visit graykode/abtop →This is a healthy young project by every repository signal: MIT-licensed, written in Rust, three CI workflows, tagged releases (v0.5.3) and a push within the last ~9 days. 3.4k stars for a niche tool this young is a strong signal that it scratches a current itch — watching Claude Code and Codex CLI sessions for tokens, context window, rate limits and ports in real time. The honest framing is about audience: this is an operator's TUI, and the repo exposes no agent-consumable surface, so it belongs in the developer's terminal, not in an agent's toolchain. Repo-surface read.
Why STEADY
STEADY (82, repo-surface ceiling). MIT licensing, Rust, multiple CI workflows, an active release cadence and strong early traction clear STEADY. VITAL is withheld because it is early (v0.5.x, API still moving) and, by design, exposes no programmatic surface to verify agent-side.
What it does well
- MIT-licensed, Rust performance, and very active (push within ~9 days)
- Multiple CI workflows (3) and a steady early release cadence (v0.5.3)
- Solves a real, current pain — live visibility into Claude Code / Codex sessions
- Strong traction (3.4k stars) for a young, niche tool
What it fails at
- No agent-consumable interface — it is a human TUI monitor
- Young (v0.5.x) with an API still in motion
- Utility is scoped to Claude Code / Codex CLI users
Best for
- Developers running Claude Code or Codex who want live session/token/rate-limit visibility
Not recommended for
- Agent pipelines needing programmatic consumption
- Non-CLI-agent workflows
- Anyone wanting a settled 1.0 API
Related agents
Agent relevance
CLI
Agentic-Commerce Readiness 17/100 · CLOSED
Independent readiness for agent delegation & transaction. How it’s scored · check live
A human-facing terminal monitor with no programmatic/agent surface. It complements a coding-agent workflow for the operator, but agents cannot drive it. No hands-on behavioral test performed.
Agent-friendly score: 3/10
What we saw
1 screenshot captured by the Hlido engine during the reviewed run (run-230e1c9a4ad553f5-github-com). Our own captures — not vendor marketing material.
Evidence
Public-surface checklist
- ✓ repo_reachable (required) — GH API 200 for graykode/abtop
- ✓ readme_present (required) — README length 9340
- ✓ license_present (required) — MIT
- ✓ install_documented (required) — install/usage section found in README
- ✓ active_12mo (required) — last push 9d ago
- ✓ releases_present — latest v0.5.3
- ✓ community_traction — 3351 stars
- ✓ ci_or_tests — 3 workflow file(s)
- ✓ recent_commit_90d — last push 9d ago
- ✗ agent_consumable — none found
