AgentSkillsCN

JIRA Integration Skill

JIRA 集成技能

SKILL.md

JIRA Integration Skill

Description

Automatically use this skill when user wants to interact with JIRA.

Trigger when user says things like:

  • "Review 376 for context" or "review TTRC-376" (query ticket for context)
  • "Review 123" (any number = ticket ID, query it)
  • "Create a card for [description]"
  • "Create a ticket for [description]"
  • "Create a bug for [issue]"
  • "Create a story for [feature]"
  • "What's the status of 376?"
  • "Update the ticket to In Progress"
  • "Transition to Done"
  • "Add a comment to the ticket"
  • "I'm done, update the ticket"
  • "Close out the JIRA"

Do NOT trigger when:

  • User just references a ticket in passing without asking to review/query it
  • Discussing JIRA workflow concepts without needing actual queries

Key patterns:

  • "review [number]" → Query that TTRC ticket for context
  • "create a card/ticket" → Create new JIRA issue
  • Any 3-digit number in context of work → Likely a ticket ID (TTRC-XXX)

JIRA Instance Details (HARDCODED - use these directly)

code
Cloud ID:    f04decff-2283-43f1-8e60-008935b3d794
Project Key: TTRC
Project ID:  10035
Site URL:    https://ajwolfe37.atlassian.net

Issue Type IDs (use these exact IDs when creating)

TypeIDWhen to Use
Story10004DEFAULT - all dev work
Bug10046Something is broken
Epic10000Explicit request to group stories
Task10044NEVER USE
Sub-task10045NEVER USE

Status Workflow

BacklogIn ProgressIn ReviewReady for TestDone


Purpose

Handle JIRA operations via subagent to isolate 20-30K token context cost. JIRA MCP tools return full ticket dumps - this skill processes internally and returns only summaries.


Instructions

Launch a Task tool (general-purpose subagent) with the JIRA operation. The subagent will:

  1. Execute the JIRA operation using MCP tools
  2. Process the full response internally (absorbing the 20-30K tokens)
  3. Return only a concise summary to main conversation

Subagent Prompt Template

code
Execute JIRA operation: [DESCRIBE WHAT USER WANTS]

**USE THESE VALUES DIRECTLY (already known):**
- Cloud ID: f04decff-2283-43f1-8e60-008935b3d794
- Project Key: TTRC
- Story issue type ID: 10004
- Bug issue type ID: 10046
- Epic issue type ID: 10000

**MCP Tools:**
- mcp__atlassian__getJiraIssue(cloudId, issueIdOrKey) - Query ticket
- mcp__atlassian__transitionJiraIssue(cloudId, issueIdOrKey, transition) - Change status
- mcp__atlassian__addCommentToJiraIssue(cloudId, issueIdOrKey, commentBody) - Add comment
- mcp__atlassian__createJiraIssue(cloudId, projectKey, issueTypeName, summary, description) - Create issue
- mcp__atlassian__editJiraIssue(cloudId, issueIdOrKey, fields) - Update fields
- mcp__atlassian__searchJiraIssuesUsingJql(cloudId, jql) - Search

**Example - Query ticket:**
mcp__atlassian__getJiraIssue(cloudId="f04decff-2283-43f1-8e60-008935b3d794", issueIdOrKey="TTRC-376")

**Example - Create Story:**
mcp__atlassian__createJiraIssue(
  cloudId="f04decff-2283-43f1-8e60-008935b3d794",
  projectKey="TTRC",
  issueTypeName="Story",
  summary="Add feature X",
  description="Description here"
)

Return ONLY a summary in this format:
✅ TTRC-XXX: [Title]
Type: [Story/Bug/Epic] | Status: [Status]
Labels: [labels if any]
[Action taken or key finding]

Do NOT return full descriptions, comment history, or raw API responses.

Issue Type Rules (STRICT)

When creating issues:

Bug

Use ONLY when something existing is broken.

  • Keywords: fix, broken, error, wrong, not working, crash, fail
  • Example: "Duplicate articles appearing" → Bug

Story (DEFAULT)

Use for ALL other work. When in doubt, use Story.

  • Keywords: add, implement, create, build, enhance, improve, refactor, update
  • Example: "Add entity-based dedup" → Story
  • Example: "Improve clustering" → Story
  • Example: "Update docs" → Story

Epic

ONLY when user explicitly says "epic" or "initiative".

  • Must be explicit request to group Stories
  • Example: "Create an epic for Clustering V2" → Epic

NEVER USE

  • Task
  • Sub-task

Labels

Apply based on work area:

  • clustering - Story clustering logic
  • rss - RSS pipeline
  • ui - Frontend/UI work
  • infra - Infrastructure/DevOps
  • security - Security hardening
  • docs - Documentation

Context Savings

Main conversation: ~100 tokens (summary only) Subagent absorbs: 20-30K tokens (full JIRA response) Savings: 99.5%