dongkeren

8 ms·

1.
Ask HN: Have your coding agents finished work you no longer wanted?
2 points by dongkeren 24d ago|0 comments
2.
by dongkeren 27d ago
I think PR doesn't mean actual work load, but this level of throughput is still impressive. I guess there was something unusual in the workflow.
3.
by dongkeren 29d ago
That's something I have been working on because I want to decouple work from the agent. In my opinion, work should be treated as a first-class object. A long running work could have many attempts, failures and retries should be aggrega
4.
by dongkeren 29d ago
It makes sense to take cost per task more important than cost per token, but for long running tasks, are more appropriate indicator could be "cost per accepted outcome". Because for a task, it is hard to say: What is the boundary
5.
by dongkeren 29d ago
I am wondering how it solves the security issues: I saw opt-in, file permissions and SSRF in README, but I do not see: domain allowlist; human approval before submiting/deleting; persistent audit record after operations; how to revoke
6.
by dongkeren 1mo ago
You'd better take harness and provider as different parts. In your case, I think the major issue is OpenCode go provier, not necessariy OpenCode harness. If you just want to have an alternative provider, I would recommend OpenRouter,
7.
Show HN: Kungfu – Keep coding-agent work alive across sessions and handoffs(github.com)
1 points by dongkeren 1mo ago|0 comments
8.
Show HN: Kungfu – One human, agents, and 3,467 merged PRs in 30 days(kungfu.tech)
1 points by dongkeren 1mo ago|0 comments