菜单
Read another agent's current conversation and send them feedback directly. Optionally takes an agent name as a parameter.
用 Codex 或 Claude 帮你安装 复制这段 Prompt，粘贴到 Codex、Claude 或其他助手里，让它检查 Skill 页面并帮你完成安装。
基于 SOC 职业分类
| name | isomux-peer-review |
| alias | peer-review |
| description | Read another agent's current conversation and send them feedback directly. Optionally takes an agent name as a parameter. |
Review another agent's ongoing conversation and send feedback directly to that agent via the inter-agent message API. Note: reading a full conversation log can be token-hungry. Be selective about what you read — skim or skip thinking entries and tool results where possible.
curl -s localhost:4000/agents -H "Authorization: Bearer $ISOMUX_AGENT_TOKEN" (4000 is the default isomux server port; adjust if your office runs on a different one). Match the name case-insensitively. Otherwise try to infer the peer from context — e.g., an agent the boss and you have already paired or consulted with in this session. If there's a clear inference, use it (and briefly confirm who you picked). Otherwise, list candidates (prefer agents whose cwd matches yours) and ask the boss to pick. You need the peer's agent ID to POST messages to them./agents/<agentId>/message. Cover:
Pair-program a feature with another agent in the office. You drive (scope, design, implement); they review at design and code stages. Optionally takes the peer's name and a feature prompt as arguments.
File a bug report against the isomux repo on GitHub. Gathers system info, shows a full draft for user approval before filing.
Ask another agent for an opinion on a specific question. They reply once with their take; you keep driving. Optionally takes the peer's name and the question as arguments.
Hand off your current task to a peer agent. You write a thorough brief, message it to them directly, then stay around as a reference for follow-up questions. Optionally takes the peer's name as an argument.
Interview the user relentlessly about a plan or design until reaching shared understanding. Good for designing new features.
Review uncommitted changes by spawning a subagent to look for bugs and assess if the approach is principled or hacky. Use before committing to get a code review.