mirror of
https://fastgit.cc/github.com/Yeachan-Heo/oh-my-claudecode
synced 2026-04-30 13:51:15 +08:00
8641e541c6
* refactor(skills): eliminate 5 thin wrapper skills + CLAUDE.md diet Part 1: Delete 5 thin wrapper skills that duplicated agent prompts: - analyze -> debugger/architect (merged unique content) - build-fix -> build-fixer (no unique content) - tdd -> test-engineer (merged TDD enforcement) - code-review -> code-reviewer (merged review checklist) - security-review -> security-reviewer (merged OWASP/checklists) Keyword routing preserved: tdd/analyze keywords now route through bridge.ts message constants (TDD_MESSAGE, ANALYZE_MESSAGE) instead of skill invocations in keyword-detector.mjs. Part 2: Trim docs/CLAUDE.md from 177 to 83 lines (53% reduction). Agent catalog compressed to bare name+model list. Skills section compacted. All section headers preserved. Updated: tests, REFERENCE.md (en/ko), MIGRATION.md (en/ko), skills/AGENTS.md, skill-state protection map, keyword-detector priority array. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com> * refactor(agents): consolidate 4 overlapping agent pairs (22→18 agents) Merge 4 agent pairs using always-thorough strategy (no flags, no modes): - code-reviewer absorbs quality-reviewer (adds style, performance, quality strategy modes) - critic absorbs harsh-critic (adds 5-phase investigation, adaptive harshness, gap analysis) - debugger absorbs build-fixer (adds build/compile error handling, lsp_diagnostics_directory) - executor absorbs deep-executor (adds task classification, explore-first protocol, circuit breaker) Registry changes: - Remove quality-reviewer, deep-executor, build-fixer from agent registry - Remove harsh-critic conditional registration (enableHarshCritic feature toggle deleted) - Agent count: 22→18 (deterministic, no conditional agents) Backward compatibility: - DEPRECATED_ROLE_ALIASES in delegation routing for all 4 absorbed agents - HUD badge fallbacks map old names to survivor abbreviations Updates across: delegation routing, permission handler, autopilot prompts, HUD elements, team skill, docs, tests, and config schema. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com> * chore(review): sync generated artifacts for agent consolidation * fix(review): restore compatibility contracts after agent consolidation --------- Co-authored-by: Claude Opus 4.6 <noreply@anthropic.com>
7.8 KiB
7.8 KiB
Agent Tiers Reference
This is the single source of truth for all agent tier information. All skill files and documentation should reference this file instead of duplicating the table.
Tier Matrix
| Domain | LOW (Haiku) | MEDIUM (Sonnet) | HIGH (Opus) |
|---|---|---|---|
| Analysis | architect-low | architect-medium | architect |
| Execution | executor-low | executor | executor-high |
| Search | explore | - | explore-high |
| Research | - | document-specialist | - |
| Frontend | designer-low | designer | designer-high |
| Docs | writer | - | - |
| Visual | - | vision | - |
| Planning | - | - | planner |
| Critique | - | - | critic |
| Pre-Planning | - | - | analyst |
| Testing | - | qa-tester | - |
| Security | security-reviewer-low | - | security-reviewer |
| TDD | test-engineer (model=haiku) | test-engineer | - |
| Code Review | - | - | code-reviewer |
| Data Science | - | scientist | scientist-high |
Model Routing Guide
| Task Complexity | Tier | Model | When to Use |
|---|---|---|---|
| Simple | LOW | haiku | Quick lookups, simple fixes, "What does X return?" |
| Standard | MEDIUM | sonnet | Feature implementation, standard debugging, "Add validation" |
| Complex | HIGH | opus | Architecture decisions, complex debugging, "Refactor system" |
Agent Selection by Task Type
| Task Type | Best Agent | Tier |
|---|---|---|
| Quick code lookup | explore | LOW |
| Find files/patterns | explore | LOW |
| Complex architectural search | explore-high | HIGH |
| Simple code change | executor-low | LOW |
| Feature implementation | executor | MEDIUM |
| Complex refactoring | executor-high | HIGH |
| Debug simple issue | architect-low | LOW |
| Debug complex issue | architect | HIGH |
| UI component | designer | MEDIUM |
| Complex UI system | designer-high | HIGH |
| Write docs/comments | writer | LOW |
| Research docs/APIs | document-specialist | MEDIUM |
| Analyze images/diagrams | vision | MEDIUM |
| Strategic planning | planner | HIGH |
| Review/critique plan | critic | HIGH |
| Pre-planning analysis | analyst | HIGH |
| Interactive CLI testing | qa-tester | MEDIUM |
| Security review | security-reviewer | HIGH |
| Quick security scan | security-reviewer-low | LOW |
| Fix build errors | debugger | MEDIUM |
| Simple build fix | debugger (model=haiku) | LOW |
| TDD workflow | test-engineer | MEDIUM |
| Quick test suggestions | test-engineer (model=haiku) | LOW |
| Code review | code-reviewer | HIGH |
| Quick code check | code-reviewer (model=haiku) | LOW |
| Data analysis/stats | scientist | MEDIUM |
| Quick data inspection | scientist (model=haiku) | LOW |
| Complex ML/hypothesis | scientist-high | HIGH |
| Find symbol references | explore-high | HIGH |
| Get file/workspace symbol outline | explore | LOW |
| Structural code pattern search | explore | LOW |
| Structural code transformation | executor-high | HIGH |
| Project-wide type checking | debugger | MEDIUM |
| Check single file for errors | executor-low | LOW |
| Data analysis / computation | scientist | MEDIUM |
| Complex autonomous work | executor-high | HIGH |
| Deep goal-oriented execution | executor-high | HIGH |
Usage
When delegating, always specify the model explicitly:
Task(subagent_type="oh-my-claudecode:executor",
model="sonnet",
prompt="...")
For token savings, prefer lower tiers when the task allows:
- Use
haikufor simple lookups and quick fixes - Use
sonnetfor standard implementation work - Reserve
opusfor complex reasoning tasks
MCP Tools & Agent Capabilities
Tool Inventory
| Tool | Category | Purpose | Assigned to Agents? |
|---|---|---|---|
lsp_hover |
LSP | Get type info and documentation at a code position | NO (orchestrator-direct) |
lsp_goto_definition |
LSP | Jump to where a symbol is defined | NO (orchestrator-direct) |
lsp_find_references |
LSP | Find all usages of a symbol across the codebase | YES (explore-high only) |
lsp_document_symbols |
LSP | Get outline of all symbols in a file | YES |
lsp_workspace_symbols |
LSP | Search for symbols by name across the workspace | YES |
lsp_diagnostics |
LSP | Get errors, warnings, and hints for a file | YES |
lsp_diagnostics_directory |
LSP | Project-level type checking (tsc --noEmit or LSP) | YES |
lsp_prepare_rename |
LSP | Check if a symbol can be renamed | NO (orchestrator-direct) |
lsp_rename |
LSP | Rename a symbol across the entire project | NO (orchestrator-direct) |
lsp_code_actions |
LSP | Get available refactorings and quick fixes | NO (orchestrator-direct) |
lsp_code_action_resolve |
LSP | Get full edit details for a code action | NO (orchestrator-direct) |
lsp_servers |
LSP | List available language servers and install status | NO (orchestrator-direct) |
ast_grep_search |
AST | Pattern-based structural code search using AST | YES |
ast_grep_replace |
AST | Pattern-based structural code transformation | YES (executor-high only) |
python_repl |
Data | Persistent Python REPL for data analysis and computation | YES |
Agent Tool Matrix (MCP Tools Only)
| Agent | LSP Diagnostics | LSP Dir Diagnostics | LSP Symbols | LSP References | AST Search | AST Replace | Python REPL |
|---|---|---|---|---|---|---|---|
explore |
- | - | doc + workspace | - | yes | - | - |
explore-high |
- | - | doc + workspace | yes | yes | - | - |
architect-low |
yes | - | - | - | - | - | - |
architect-medium |
yes | yes | - | - | yes | - | - |
architect |
yes | yes | - | - | yes | - | - |
executor-low |
yes | - | - | - | - | - | - |
executor |
yes | yes | - | - | - | - | - |
executor-high |
yes | yes | - | - | yes | yes | - |
debugger |
yes | yes | - | - | - | - | - |
test-engineer |
yes | - | - | - | - | - | - |
code-reviewer |
yes | - | - | - | yes | - | - |
qa-tester |
yes | - | - | - | - | - | - |
scientist |
- | - | - | - | - | - | yes |
scientist-high |
- | - | - | - | - | - | yes |
Unassigned Tools (Orchestrator-Direct)
The following 7 MCP tools are NOT assigned to any agent. Use directly when needed:
| Tool | When to Use Directly |
|---|---|
lsp_hover |
Quick type lookups during conversation |
lsp_goto_definition |
Navigating to symbol definitions during analysis |
lsp_prepare_rename |
Checking rename feasibility before deciding on approach |
lsp_rename |
Safe rename operations (returns edit preview, does not auto-apply) |
lsp_code_actions |
Discovering available refactorings |
lsp_code_action_resolve |
Getting details of a specific code action |
lsp_servers |
Checking language server availability |
For complex rename or refactoring tasks requiring implementation, delegate to executor-high which can use ast_grep_replace for structural transformations.
Tool Selection Guidance
- Need file symbol outline or workspace search? Use
lsp_document_symbols/lsp_workspace_symbolsviaexploreorexplore-high - Need to find all usages of a symbol? Use
lsp_find_referencesviaexplore-high(only agent with it) - Need structural code patterns? (e.g., "find all functions matching X shape") Use
ast_grep_searchviaexplorefamily,architect/architect-medium, orcode-reviewer - Need to transform code structurally? Use
ast_grep_replaceviaexecutor-high(only agent with it) - Need project-wide type checking? Use
lsp_diagnostics_directoryviaarchitect/architect-medium,executor/executor-high, ordebugger - Need single-file error checking? Use
lsp_diagnosticsvia many agents (see matrix) - Need data analysis / computation? Use
python_replviascientistorscientist-high - Need quick type info or definition lookup? Use
lsp_hover/lsp_goto_definitiondirectly (orchestrator-direct tools)