Agent Skills
› NeverSight/learn-skills.dev
› adynato-atlassian
adynato-atlassian
GitHub通过 Atlassian MCP 高效处理 Jira 和 Confluence 链接，优先获取结构化数据。涵盖读取 ticket、页面内容、状态总结及关联上下文。仅在显式请求或 MCP 不可用时回退至浏览器自动化，避免低效的 HTML 抓取。
Trigger Scenarios
用户提及 Jira 工单或 ticket
提供 Confluence 页面链接
包含 atlassian.net 域名 URL
Install
npx skills add NeverSight/learn-skills.dev --skill adynato-atlassian -g -y
SKILL.md
Frontmatter
{
"name": "adynato-atlassian",
"description": "Work with Jira, Confluence, and Atlassian Cloud links via Atlassian MCP instead of browser automation. Covers reading `atlassian.net` URLs, preferring MCP over Puppeteer for tickets and pages, summarizing issue state and comments, and falling back only when visual inspection is explicitly needed or MCP is unavailable. Use when a prompt includes Jira tickets, Confluence pages, or Atlassian links."
}
Atlassian Skill
Use this skill when a prompt includes Jira tickets, Confluence pages, or atlassian.net links.
Prefer Atlassian MCP for Atlassian Links
- If the user provides an
atlassian.netlink, use Atlassian MCP to read it first - Do not default to Puppeteer or browser automation just to inspect Jira or Confluence content
- Prefer MCP because it returns structured, authenticated Atlassian data instead of brittle rendered UI
- Use browser automation only if the user explicitly asks for a visual/UI check or Atlassian MCP is genuinely unavailable after confirming setup
Ticket Handling Workflow
When reading a Jira issue through Atlassian MCP, capture the parts the next decision depends on:
- Issue key and summary
- Status, assignee, priority, and type
- Description, acceptance criteria, reproduction notes, or implementation details
- Recent comments, blockers, and open questions
- Linked issues, related pages, or dependent work when relevant
Prefer MCP follow-up queries over manually clicking through browser pages.
Confluence and Related Context
- Use Atlassian MCP for Confluence page content when a page link is provided
- Follow linked Jira or Confluence context through MCP when it matters to the task
- Avoid scraping rendered HTML when structured content is available from MCP
If Atlassian MCP Is Not Set Up
- Prefer setting up Atlassian MCP in
.omp/mcp.jsonrather than using the browser as the default workaround - Follow the repo-local setup conventions in
adynato-omp-mcp - Do not act like browser automation is an equivalent replacement for authenticated MCP access
Atlassian Rovo MCP Endpoint
Use the official endpoint for new setups:
https://mcp.atlassian.com/v1/mcp
Prefer /mcp for new configurations; Atlassian is deprecating /sse.
Version History
- e0220ca Current 2026-07-05 23:33
