atlassian-rovo-mcp
Set up, authenticate, verify, and use the Atlassian Rovo remote MCP server for Jira, Confluence, Bitbucket, Compass, and Teamwork Graph work. Use when Codex or Claude needs Atlassian access, needs the official Rovo MCP server configured, or when replacing former local Atlassian CLI workflows.
- License 许可
- MIT
- 最近更新
- 2026-07-19
- Stars
- ★ 0
- Forks
- 0
01 前言 Front Matter
- name
- atlassian-rovo-mcp
- description
- Set up, authenticate, verify, and use the Atlassian Rovo remote MCP server for Jira, Confluence, Bitbucket, Compass, and Teamwork Graph work. Use when Codex or Claude needs Atlassian access, needs the official Rovo MCP server configured, or when replacing former local Atlassian CLI workflows.
02 资源文件 1
- 参考文档agents/openai.yaml437 B
03 SKILL.md 正文
Atlassian Rovo MCP
Overview
Use Atlassian Rovo MCP as the default Atlassian integration for this repository. The former local Atlassian CLI path is no longer the adapter for Jira or Confluence project-memory integrations.
Server
- MCP server name:
atlassian - MCP server URL:
https://mcp.atlassian.com/v1/mcp/authv2 - Official setup guide:
https://support.atlassian.com/atlassian-rovo-mcp-server/docs/getting-started-with-the-atlassian-remote-mcp-server/
Setup
For Codex:
codex mcp add atlassian --url https://mcp.atlassian.com/v1/mcp/authv2
codex mcp login atlassian
codex mcp get atlassian
For Claude Code:
claude mcp add --transport http atlassian https://mcp.atlassian.com/v1/mcp/authv2
Then open a Claude Code session and run /mcp to authenticate.
For JSON-based MCP clients:
{
"mcpServers": {
"atlassian": {
"url": "https://mcp.atlassian.com/v1/mcp/authv2"
}
}
}
If the server already exists, verify the configured URL before changing anything. Remove and re-add the server only when the existing entry points somewhere else.
Use
- Prefer the
atlassianMCP server for Jira, Confluence, Bitbucket, Compass, and Teamwork Graph work. - Inspect the MCP tools exposed by the current harness before acting; do not invent tool names.
- Use
$jira-rovo-mcpfor Jira project-memory integrations and$confluence-rovo-mcpfor Confluence project-memory integrations. - Treat local project memory as canonical. Atlassian is a visibility and collaboration surface.
- Ask before creating or updating external Atlassian objects unless the user explicitly requested the action or the project integration policy permits auto-publishing.
- Never publish secrets, credentials, private notes, raw logs, bulky output, or sensitive local paths.
- If MCP is missing, unauthenticated, or unavailable on the current node, save a local draft and record the publish status as skipped for that node.
Do not rely on local Atlassian CLI skills. If MCP is unavailable, follow the relevant project integration policy and keep a local draft.
04 同仓库更多 Skills
status
Use only when the user invokes /status or explicitly asks to update the project memory context card, blocker, latest result, or next action. Do not use for ordinary git status unless the user asks to log project status in memory.
scratch
Use only when the user invokes /scratch or explicitly asks to capture uncertain, in-flight, project-local working notes that are not ready for durable memory.
run
Use only when the user invokes /run or explicitly asks to record or update an experiment, evaluation, benchmark, or long-running job in project memory. Do not use for executing shell commands, running tests, or starting ordinary programs unless the user asks to log that run in memory.
remember
Use only when the user invokes /remember or asks to record project activity, a run, decision, learning, status change, blocker, next action, or compact daily memory. Do not use for unrelated reminders or task scheduling.
organise-scratch
Use only when the user invokes /organise-scratch or asks to route project-local scratch notes into durable project memory files.
log
Use only when the user invokes /log or explicitly asks to record session work or a meaningful project change in project memory. Do not use merely to inspect application logs.