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。
直接命令不会经过审查 Prompt;运行前请先检查来源。
npx skills add https://github.com/Arvo-AI/aurora --skill slack命令会保持在同一行。复制前请横向滚动并检查完整内容。
想先保存到本地?可下载 SkillsMP 当前能够提供的文件。
正在显示 SKILL.md
| name | slack |
| id | slack |
| description | Slack integration tools for reading channel messages and thread replies |
| category | communication |
| connection_check | {"method":"is_connected_function","module":"chat.backend.agent.tools.slack_tool","function":"is_slack_connected"} |
| tools | ["list_slack_channels","get_channel_history","get_thread_replies"] |
| index | Slack messaging -- list channels, read messages, read threads |
| rca_priority | 50 |
| metadata | {"author":"aurora","version":"1.0"} |
Read-only tools for searching Slack conversations. Used during postmortem generation to gather human context (deployment decisions, communication gaps, resolution steps) and during interactive chat for incident investigation.
list_slack_channels()Returns all channels the bot can access: id, name, topic, purpose, member count. Use channel names and topics to identify relevant channels for the incident (look for service names, "incident", "oncall", "alerts").
get_channel_history(channel_id, oldest?, latest?, limit?)Fetch messages from a channel. Scope with oldest/latest (ISO 8601) to the incident time window. Returns message text, timestamps, user IDs, and thread metadata (reply_count, thread_ts).
get_thread_replies(channel_id, thread_ts, limit?)Fetch replies in a thread. Use when a message has reply_count > 0 and looks relevant.
list_slack_channels — scan names/topics for the affected service or "incident"/"oncall" keywordsget_channel_history on the most relevant channels, scoped to the incident time windowreply_count > 0 and looks relevant, call get_thread_replies for full context基于 SOC 职业分类