跳到主要内容 / Skip to main content
🌐V2EX260v2ex.com

[AI Agent 智能体] 话说在 Codex 或者 Claude Code 里开 Plan 的时候,有人会给 agent veto 用户决策的想法吗?

首次上榜 2026-09-29

摘要

一位开发者在 V2EX 分享自己在 Codex、Claude Code 的 Plan 阶段主动授权 agent 否决自己决策的做法:当 Plan 的多个问题互相耦合、自己的偏保守选择可能让后续问题难度跳档时,他会给出偏好并附加「若有充分证据可以 veto 我」,agent 随后判断按他的思路整个 Plan 极大概率会变题并予以否决。他也提出疑问,在单人和单 agent 的场景下 veto 意义有限,如果放进多个 agent 互相通信的 loop 里,这个权限是否会更有价值。

相关推荐

tt-a1i/archify

Turn any idea, plan, or codebase into a beautiful interactive diagram. An agent skill for Claude Code, Codex, and more.

agentbeautifulclaudecodebasecodex

affaan-m/ECC

The agent harness performance optimization system. Skills, instincts, memory, security, and research-first development for Claude Code, Codex, Opencode, Cursor and beyond.

K-Dense-AI/scientific-agent-skills

Turn any AI agent into an AI Scientist. The #1 Agent Skills library for science, used by 250,000+ scientists worldwide. 177 ready-to-use validated skills plus 100+ scientific databases covering biology, chemistry, medicine, and drug discovery. Compatible with Cursor, Claude Code, Codex, Pi, Antigravity, and the open Agent Skills standard.

agentaiantigravitybiologychemistry