using-confluence

原文涵盖的主题

• The wiki
• What the tools cannot read
• Macros and mentions
• Searching efficiently
• Writing to Confluence
• Revising a page
• Never

The project's own facts are in /wiki.md, being the directory python3 "${CLAUDEPLUGINROOT}/scripts/profile.py" --standard prints: where its Confluence is and how a page is linked, its spaces and what each holds and which parts to read, the pages and page IDs a skill reads by name, what to skip, the pages the monthly revision watches, its queries, and the values a Jira macro and a mention take there. Read it before the first Confluence call. When --standard exits 1 ending "the repository has no standard", as outside the project's repositories, or the standard holds no wiki.md, say so and ask which project's standard to read, as / , whose standards/wiki.md is then read from that repository through GitHub, or whether to go on without one, with what the person pastes; any other exit 1 is a refusal: show its message and ask the same. With no file read, find the spaces the person names by search and ask once for what search does not give. Reach Confluence through the Atlassian MCP server, resolved with ToolSearch on confluence. The server may run read-write; the skills read it, and write only as "Writing to Confluence" below sets out. A server that is not connected, or a call that fails,

这里只展示节选，请以完整原文和当前要求为准。