jira
Use when the user asks to interact with Jira — search issues, read or update tickets, transition status, add comments, list projects, or run JQL. Routes operations through the Atlassian Rovo MCP server (mcp__atlassian__* tools).
- License 许可
- MIT
- 最近更新
- 2026-11-15
- Stars
- ★ 0
- Forks
- 1
01 前言 Front Matter
- name
- jira
- triggers
- [ "user" ]
- description
- Use when the user asks to interact with Jira — search issues, read or update tickets, transition status, add comments, list projects, or run JQL. Routes operations through the Atlassian Rovo MCP server (mcp__atlassian__* tools).
- allowed-tools
- [ "mcp_call_tool", "mcp_list_tools", "ask_user_question" ]
03 SKILL.md 正文
Jira operations via Atlassian Rovo MCP
This skill routes Jira interactions through the Atlassian Rovo MCP server (atlassian), configured at user scope via devin mcp add atlassian -s user https://mcp.atlassian.com/v1/mcp/authv2 + devin mcp login atlassian. The config lives in %APPDATA%\devin\mcp_config.json so it works in any project directory. All operations respect the authenticated user's Jira permissions and scopes (read:jira-work, write:jira-work).
Prerequisites
- MCP server
atlassianmust be configured (devin mcp listshows it). - OAuth completed (
devin mcp login atlassianif expired). - Jira Cloud only (the official Rovo MCP does not support Server/Data Center).
Known site
Read site values from data/bundle-integrations.json (jira.site, jira.cloud_id) or the environment variables JIRA_SITE and JIRA_CLOUD_ID. Set a default project in data/bundle-integrations.json (jira.project) or via JIRA_PROJECT.
- cloudId:
${JIRA_CLOUD_ID}(fallback:data/bundle-integrations.jsonjira.cloud_id) - site:
${JIRA_SITE}(fallback:data/bundle-integrations.jsonjira.site) - project:
${JIRA_PROJECT}(fallback:data/bundle-integrations.jsonjira.project)
Pass cloudId and site to every Jira tool call. If a new site is added, run getAccessibleAtlassianResources and update the integration config.
Tool namespace
All Jira tools are exposed under the atlassian MCP server as mcp__atlassian__<tool>. Call them via mcp_call_tool with server_name: "atlassian".
To discover the current tool set at runtime:
mcp_list_tools -> server_name: "atlassian"
Core operations
1. Discover available tools
mcp_list_tools(server_name="atlassian")
Use this first in a new session to confirm the exact tool names (Atlassian may add/rename tools between releases).
2. List visible projects
mcp_call_tool(
server_name="atlassian",
tool_name="getVisibleJiraProjects",
arguments={}
)
3. JQL search
mcp_call_tool(
server_name="atlassian",
tool_name="searchJiraIssuesUsingJql",
arguments={"jql": "project = PROJ AND status = Open ORDER BY created DESC", "maxResults": 25}
)
Common JQL patterns:
- My open issues:
assignee = currentUser() AND statusCategory != Done - Sprint issues:
project = PROJ AND sprint in openSprints() - Recently updated:
project = PROJ AND updated >= -7d ORDER BY updated DESC - Blocked:
project = PROJ AND status = Blocked
4. Get a single issue
mcp_call_tool(
server_name="atlassian",
tool_name="getJiraIssue",
arguments={"issueKey": "PROJ-123"}
)
5. Create an issue
mcp_call_tool(
server_name="atlassian",
tool_name="createJiraIssue",
arguments={
"projectId": "<numeric id from getVisibleJiraProjects>",
"issueTypeId": "<id from getJiraProjectIssueTypesMetadata>",
"summary": "Short title",
"description": "Markdown or ADF description"
}
)
Always fetch project + issue type metadata before creating — IDs are numeric, not string keys.
6. Edit an issue
mcp_call_tool(
server_name="atlassian",
tool_name="editJiraIssue",
arguments={"issueKey": "PROJ-123", "summary": "Updated title"}
)
7. Transition status
mcp_call_tool(
server_name="atlassian",
tool_name="transitionJiraIssue",
arguments={"issueKey": "PROJ-123", "transitionId": "<id>"}
)
Transition IDs are workflow-specific. Fetch them via the issue's available transitions before calling.
8. Add a comment
mcp_call_tool(
server_name="atlassian",
tool_name="addCommentToJiraIssue",
arguments={"issueKey": "PROJ-123", "comment": "Comment text in Markdown"}
)
9. Link two issues
mcp_call_tool(
server_name="atlassian",
tool_name="createIssueLink",
arguments={
"issueKey": "PROJ-123",
"linkedIssueKey": "PROJ-456",
"linkTypeId": "<id from getIssueLinkTypes>"
}
)
10. Worklog (time tracking)
mcp_call_tool(
server_name="atlassian",
tool_name="addWorklogToJiraIssue",
arguments={"issueKey": "PROJ-123", "timeSpent": "2h", "comment": "Refactored auth module"}
)
Workflow patterns
Triage an issue end-to-end
getJiraIssue(issueKey)— read full contextsearchJiraIssuesUsingJql— find related/blocked issues- Read the codebase (
grep,read) to understand the change - Implement the fix
addCommentToJiraIssue— post a summary of what was changedtransitionJiraIssue— move to Review/Done
Create a ticket from a finding
getVisibleJiraProjects— confirm project IDgetJiraProjectIssueTypesMetadata— pick issue typecreateJiraIssue— with summary + description referencing the code location- Confirm the created key to the user
Safety rules
- Never transition or delete without user confirmation. Use
ask_user_questionbefore destructive state changes. - Write operations need explicit user intent. Don't auto-comment or auto-transition because a task "looks done".
- Rate limits: Free 500/hr, Standard 1k/hr, Premium/Enterprise up to 10k/hr. For bulk operations, batch and pause on 429.
- No secrets in arguments. The MCP server holds the OAuth token; never pass tokens, passwords, or API keys in tool arguments.
When this skill does NOT apply
- Jira Server/Data Center: the official Rovo MCP is Cloud-only. Use
sooperset/mcp-atlassian(self-hosted) and a different skill. - Confluence-only operations: the same
atlassianserver exposes Confluence tools, but prefer a dedicatedconfluenceskill if one exists. - Bulk exports / reporting: the MCP is not designed for large dumps; use the REST API directly for those.
Cross-skills
- Use
intakewhen moving issues through a triage state machine. - Use
planningto turn Jira issues into specs or tickets. - Use
executiononce a Jira ticket is ready for code changes. - If you only need the Atlassian MCP for this task, use
mcp-governanceto avoid leaving it active.
Troubleshooting
mcp_list_toolsreturns empty: rundevin mcp login atlassianto refresh the OAuth token.- 403 / permission denied: the authenticated user lacks the Jira project permission; check with the Jira admin, not Devin.
- 429 rate limit: pause for 60s, reduce batch size, or upgrade the Atlassian tier.
- Tool name not found: Atlassian renames tools between releases; always call
mcp_list_toolsfirst and use the exact returned name.
04 同仓库更多 Skills
montador-de-ideias
Use when assembling, expanding, or elevating a raw project idea or initial prompt into an ultra-detailed, first-principles dossier and compiled master prompt with journalistic OSINT market research, Tech Lead and Staff Architect justifications, atomic and modular hardening/optimization, and high-impact contact surfaces.
git-github-adaptive-engineering
Use when any task touches git history, GitHub issues/PRs/releases/CI, repo governance, or branch and merge decisions. Runs the Fase Zero situational-awareness battery, picks one of five delivery flows via the adaptive decision engine, and enforces commit/PR information-density standards. Orchestrates the gh, git-workflows, github-actions, pr, and debug/finish skills.
adaptive-agenda-calendar-routine-engine
Use when planning or executing work across hours, days, weeks, or sprints. Turns the calendar into a strict time-blocked agenda (HH:MM-HH:MM blocks carrying a routine profile, expected inputs, and a handoff output artifact), runs the 5-stage cognitive flow (Radar and Hunt -> Triage plus Steve Jobs conception -> workbench manifest -> >=32-cycle loop -> Close plus Open-Minded Scrum), and unfolds macro calendar milestones into micro agenda blocks through extensions/gcal/wrapper.py.
demo
Use when demonstrating a skill in a fixture project.
duplicate-skill
Use when testing duplicate skill name detection.
duplicate-skill
Use when testing duplicate skill name detection.