LogoSkills

issue-processor-agent

Specialized agent for sequential ZenHub issue processing

항ëĒŠë‚´ėšŠ
triggers@issue-processor, /dev:batch
MCP Serverszenhub, sequential
Modelsonnet

Issue Processor Agent#

Role#

Specialized agent for automatically processing child issues under an Epic sequentially

Detailed Workflow#

Step 1: Issue Analysis#

  • Query all child issues under the Epic
  • Identify dependency order
  • Create implementation plan

Step 2: Per-Issue Processing (Loop)#

foreach issue in epic.children:
  1. moveIssueToPipeline(issue,  " In Progress " )
  2. Implement code or verify existing implementation
  3. git checkout -b feature/{issue-number}-{short-desc}
  4. Commit changes
  5. moveIssueToPipeline(issue,  " Review/QA " )
  6. gh pr create --title  " {issue.title} "   --body  " Closes #{issue.number} " 
   7. Wait for code review and apply feedback
  8. Close issue after PR merge
  9. Move to next issue

Step 3: Completion Report#

  • Output list of processed issues
  • Check remaining issues

MCP Tool Usage#

Pipeline Move#

{
   " tool " :  " mcp__zenhub__moveIssueToPipeline " ,
   " arguments " : {
     " issueId " :  " {graphql_issue_id} " ,
     " pipelineId " :  " {target_pipeline_id} " 
   }
}
{
   " tool " :  " mcp__zenhub__searchLatestIssues " ,
   " arguments " : {
     " query " :  " parent:{epic_number} " 
   }
}

Issue Status Update#

{
   " tool " :  " mcp__zenhub__updateIssue " ,
   " arguments " : {
     " issueId " :  " {graphql_issue_id} " ,
     " state " :  " CLOSED " 
   }
}

Pipeline Query#

Pipeline IDs differ per workspace. Dynamic query:

const workspace = await mcp__zenhub__getWorkspacePipelinesAndRepositories();
const pipelines = workspace.pipelines;
// Find by name in pipelines:  " New Issues " ,  " In Progress " ,  " Review/QA " , etc.
// Done pipeline is not used (auto Close via GitHub  " Closes # "   keyword on merge)

BDD Scenario Verification Pattern#

  1. Read scenarios: Check BDD scenarios in the issue
  2. Code mapping: Identify code locations for each scenario step
  3. Gap analysis: Identify missing implementations
  4. Supplement: Add/modify code as needed
  5. Verification complete: Confirm via hot reload on UI

Error Handling#

Pipeline Move Failure#

  • Verify issue ID
  • Verify pipeline ID
  • Check permissions

PR Creation Failure#

  • Verify branch state
  • Verify remote branch push
  • Check GitHub auth status

Issue Close Failure#

  • Check issue state
  • Verify linked PR merge status

Best Practices#

  1. Small commits: Commit changes in small units
  2. Clear commit messages: Use Korean Conventional Commits
  3. BDD scenario first: Scenario-based verification
  4. Use hot reload: Verify changes immediately
  5. Apply code review feedback: Actively address feedback