Hi, thanks for maintaining this harness-engineering list.
I’d like to suggest adding agenttrace:
- [agenttrace](https://github.com/luoyuctl/agenttrace) — Local-first TUI for AI coding-agent session history across Claude Code, Codex, Gemini CLI, and OpenCode. Tracks cost, tokens, time, tool failures, latency, health reports, diffs, and CI gates from local logs without requiring SaaS telemetry.
Why it fits: the list emphasizes harness infrastructure around coding agents. agenttrace helps teams audit real local agent runs after the fact, especially cost attribution, failed tool calls, and health/CI evidence from session logs.
I would normally open a PR, but the existing luoyuctl/awesome-agent-harness fork is currently tied to another upstream with the same repo name, and I am avoiding creating new repositories.
Hi, thanks for maintaining this harness-engineering list.
I’d like to suggest adding
agenttrace:Agent Knowledge & MemorynearCodeBurn, or a future observability/cost section if you split that out.Why it fits: the list emphasizes harness infrastructure around coding agents. agenttrace helps teams audit real local agent runs after the fact, especially cost attribution, failed tool calls, and health/CI evidence from session logs.
I would normally open a PR, but the existing
luoyuctl/awesome-agent-harnessfork is currently tied to another upstream with the same repo name, and I am avoiding creating new repositories.