claude/.claude/skills/atlassian/SKILL.md
Interact with Atlassian Jira and Confluence via MCP. Use when the user asks about Jira issues, Confluence pages, sprints, boards, or any Atlassian task.
npx skillsauth add felixge/dotfiles atlassianInstall this skill globally with one command. Works with Claude Code, Cursor, and Windsurf.
3 of 9 scanners reported clean
Some scanners were skipped, did not run, or reported a non-clean status. Review each row below.
Use the atlassian MCP server to interact with Jira and Confluence.
When this skill is invoked, always perform a search immediately using searchAtlassian with the user's query — don't just explain how to search or ask for clarification. The user expects results, not instructions.
searchAtlassian (searches both Jira and Confluence, no cloudId needed) — do this by defaultfetchAtlassian using the ARI (id field) from search resultssearchJiraIssuesUsingJql (requires cloudId)searchConfluenceUsingCql (requires cloudId)Most tools require a cloudId parameter. Use the site hostname: datadoghq.atlassian.net
mcporter list atlassian --schema
mcporter call atlassian.<tool_name> key=value key2=value2
# General search (both Jira and Confluence, no cloudId needed)
mcporter call atlassian.searchAtlassian query="full host profiler"
# Fetch full content using ARI from search results
mcporter call atlassian.fetchAtlassian id="ari:cloud:confluence:66c05bee-f5ff-4718-b6fc-81351e5ef659:page/3834216651"
mcporter call atlassian.fetchAtlassian id="ari:cloud:jira:66c05bee-f5ff-4718-b6fc-81351e5ef659:issue/1072997"
# Search Jira with JQL
mcporter call atlassian.searchJiraIssuesUsingJql cloudId=datadoghq.atlassian.net jql="project = PROJ AND status = Open"
# Get a specific Jira issue
mcporter call atlassian.getJiraIssue cloudId=datadoghq.atlassian.net issueIdOrKey=PROJ-123
# Search Confluence with CQL
mcporter call atlassian.searchConfluenceUsingCql cloudId=datadoghq.atlassian.net cql='title ~ "meeting notes" AND type = page'
fetchAtlassian over getConfluencePage: The getConfluencePage tool requires pageId as a string type, but mcporter parses numeric values as numbers, causing type errors. Using fetchAtlassian with ARIs from search results avoids this problem entirely.searchAtlassian is the best starting point — it doesn't need a cloudId and covers both Jira and Confluence via Rovo Search.contentFormat=markdown or responseContentFormat=markdown for readable output instead of ADF (Atlassian Document Format).mcporter auth atlassian on first use.tools
Use mcporter to call MCP servers — either from the CLI or from TypeScript scripts. Use when the user wants to list/call MCP tools, automate MCP interactions, or build scripts that talk to MCP servers.
tools
Use the jj version control system. Always use this instead of git for all version control operations including commits, diffs, logs, branches, rebasing, pushing, pulling, and any other VCS tasks.
tools
Interactively split a jj commit into multiple commits using the builtin TUI diff editor via tmux.
tools
Interact with Google Drive, Docs, Sheets, and Slides via MCP. Use when the user asks about Google Docs, Sheets, Drive files, or any Google Workspace task.