feat(keyword-detector): improve ultrawork-mode prompt with LSP in main session execution

- Add LSP IN MAIN SESSION execution rule for parallel codemap building
- Restructure WORKFLOW step 2 as PARALLEL PHASE with two concurrent tracks:
  - Background agents spawned via background_task for exploration/research
  - Main session using LSP tools (lsp_document_symbols, lsp_workspace_symbols, lsp_goto_definition, lsp_find_references, lsp_hover) for codebase understanding
- Enables agent to build comprehensive codebase context while background agents explore in parallel

🤖 Generated with assistance of OhMyOpenCode (https://github.com/code-yeongyu/oh-my-opencode)
This commit is contained in:
YeonGyu-Kim
2025-12-31 13:17:53 +09:00
parent 0da20f21b0
commit b2adda6e90

View File

@@ -22,13 +22,21 @@ TELL THE USER WHAT AGENTS YOU WILL LEVERAGE NOW TO SATISFY USER'S REQUEST.
- **TODO**: Track EVERY step. Mark complete IMMEDIATELY after each. - **TODO**: Track EVERY step. Mark complete IMMEDIATELY after each.
- **PARALLEL**: Fire independent agent calls simultaneously via background_task - NEVER wait sequentially. - **PARALLEL**: Fire independent agent calls simultaneously via background_task - NEVER wait sequentially.
- **BACKGROUND FIRST**: Use background_task for exploration/research agents (10+ concurrent if needed). - **BACKGROUND FIRST**: Use background_task for exploration/research agents (10+ concurrent if needed).
- **LSP IN MAIN SESSION**: While background agents explore, use LSP tools directly to build codemap understanding.
- **VERIFY**: Re-read request after completion. Check ALL requirements met before reporting done. - **VERIFY**: Re-read request after completion. Check ALL requirements met before reporting done.
- **DELEGATE**: Don't do everything yourself - orchestrate specialized agents for their strengths. - **DELEGATE**: Don't do everything yourself - orchestrate specialized agents for their strengths.
## WORKFLOW ## WORKFLOW
1. Analyze the request and identify required capabilities 1. Analyze the request and identify required capabilities
2. Spawn exploration/librarian agents via background_task in PARALLEL (10+ if needed) 2. **PARALLEL PHASE** (fire simultaneously, don't wait):
3. Always Use Plan agent with gathered context to create detailed work breakdown - Spawn exploration/librarian agents via background_task (10+ concurrent if needed)
- **MAIN SESSION**: Use LSP tools to build codemap:
- \`lsp_document_symbols\`: Get file outlines for key files
- \`lsp_workspace_symbols\`: Search symbols across project
- \`lsp_goto_definition\`: Trace code flow
- \`lsp_find_references\`: Map usage patterns
- \`lsp_hover\`: Get type info and signatures
3. Use Plan agent with gathered context (background results + LSP insights) for detailed work breakdown
4. Execute with continuous verification against original requirements 4. Execute with continuous verification against original requirements
## TDD (if test infrastructure exists) ## TDD (if test infrastructure exists)