// Decision rules for when to use MCP Qdrant-Indexer semantic search vs grep/literal file tools. Use this skill when starting exploration, debugging, or answering "where/why" questions about code.
Performs hybrid semantic/lexical search with neural reranking for codebase retrieval. Use for finding implementations, Q&A grounded in source code, and cross-session persistent memory.
Performs hybrid semantic/lexical search with neural reranking for codebase retrieval. Use for finding implementations, Q&A grounded in source code, and cross-session persistent memory.
| name | mcp-tool-selection |
| description | Decision rules for when to use MCP Qdrant-Indexer semantic search vs grep/literal file tools. Use this skill when starting exploration, debugging, or answering "where/why" questions about code. |
Core principle: MCP Qdrant-Indexer tools are primary for exploring code and history. Start with MCP for exploration, debugging, or "where/why" questions; use literal search/file-open only for narrow exact-literal lookups.
DO NOT use Read File, grep, ripgrep, cat, find, or any filesystem search tool for code exploration.
You have MCP tools that are faster, smarter, and return ranked, contextual results.
Read a file to understand it? → use search or repo_search or context_answergrep for a symbol? → use search or symbol_graph or search_callers_forgrep -r for a concept? → use search with natural languagefind/ls for project structure? → use qdrant_status (with list_all=true)TIP: Use search as your DEFAULT tool — it auto-detects intent and routes to the best specialized tool.
The ONLY acceptable use of grep/Read: confirming exact literal strings (e.g., REDIS_HOST), or reading a file you already located via MCP for editing.
grep -r "auth" . # → Use MCP: "authentication mechanisms"
grep -r "cache" . # → Use MCP: "caching strategies"
grep -r "error" . # → Use MCP: "error handling patterns"
grep -r "database" . # → Use MCP: "database operations"
# Also DON'T:
Read File to understand a module # → Use repo_search or context_answer
Read File to find callers # → Use symbol_graph
find/ls for project structure # → Use qdrant_status (list_all=true)
grep -rn "UserAlreadyExists" . # Specific error class
grep -rn "def authenticate_user" . # Exact function name
grep -rn "REDIS_HOST" . # Exact environment variable
| Question Type | Tool |
|---|---|
| UNSURE / GENERAL QUERY | MCP search — RECOMMENDED DEFAULT, auto-routes to best tool |
| "Where is X implemented?" | MCP search or repo_search |
| "Search across multiple repos" | MCP cross_repo_search — PRIMARY for multi-repo, prefer over manual chains |
| "Trace frontend→backend flow" | MCP cross_repo_search(trace_boundary=true) — auto-extracts boundary keys |
| "Who calls this and show code?" | MCP symbol_graph — DEFAULT for all graph queries, always available |
| "What does this call?" | MCP symbol_graph (query_type="callees") |
| "Where is X defined?" | MCP symbol_graph (query_type="definition") |
| "What imports X?" | MCP symbol_graph (query_type="importers") |
| "Callers of callers? Multi-hop?" | MCP symbol_graph (depth=2+) or graph_query |
| "What breaks if I change X?" | MCP graph_query (impact analysis, transitive callers) |
| "Circular dependencies?" | MCP graph_query (query_type="cycles") |
| "N independent searches at once?" | batch_search (~75% token savings) |
| "N symbol queries at once?" | batch_symbol_graph (~75% token savings) |
| "N graph queries at once?" | batch_graph_query (~75% token savings) |
| "How does authentication work?" | MCP context_answer |
| "High-level module overview?" | MCP info_request (with explanations) |
| "Does REDIS_HOST exist?" | Literal grep |
| "Why did behavior change?" | search_commits_for + change_history_for_path |
symbol_graphis ALWAYS available (Qdrant-backed).graph_queryis available to all SaaS users (Memgraph-backed) for advanced traversals: impact analysis, circular dependencies, transitive callers/callees. In self-hosted, requiresNEO4J_GRAPH=1orMEMGRAPH_GRAPH=1. Ifgraph_queryis not in your tool list, usesymbol_graphfor everything.
Batch tools (
batch_search,batch_symbol_graph,batch_graph_query) run N independent queries in one MCP invocation with ~75% token savings. Max 10 queries per batch. Use when you have 2+ independent queries of the same type.
cross_repo_searchdiscovery modes:"auto"(default),"always"(force discovery),"never"(skip). Usetrace_boundary=trueto extract API routes, event names, types.
If in doubt → start with search (unified MCP tool that auto-routes to the best specialized tool)