synsci-delphi

Documentation

Documentation link · transport stdio

Tools

• index_repository — Index a GitHub repository (quality_mode
• index_local_folder — Index a local directory straight from disk (private/WIP code).
• quick_index — Resolve a library name via the curated catalog and index it in one step.
• catalog_search — Resolve a library/framework name to an indexable source (works cold
• search_code — Hybrid code search (vector + BM25 + symbol + path + trigram
• search_symbols — Find functions/classes/methods by name; structural extraction for 16 languages.
• get_symbol — Full symbol details plus reconstructed source body.
• find_callers — Who calls this symbol (code-dependency graph).
• find_callees — What this symbol calls — internal + external names.
• impact_analysis — Blast radius — transitive callers.
• build_code_graph — (Re)build the call graph for a repo.
• check_freshness — Is this index stale vs remote HEAD / file re-hash.
• list_stale_sources — List indexed repos whose index has drifted.
• get_file — Retrieve file content from an indexed repo.
• get_context — Fetch a chunk plus adjacent chunks and enclosing body/siblings.
• build_context_pack — Agent-ready context pack with token budgeter and re-query planner.
• get_directory_structure — Browse repository file tree.
• analyze_repository — Deep code analysis and architecture overview.
• classify_failure — Tag a Delphi failure with a stable failure-mode code.
• index_paper — Index from arXiv URL/ID or PDF upload.
• search_papers — Hybrid paper search
• extract_quoted_evidence — Pull literal sentences from a paper that ground a claim.
• joint_retrieval — One call to paper + code + Atlas-graph hits
• get_citations — Extract citation graph.
• get_equations — Extract equations with context.
• generate_report — Generate a markdown summary report.
• compare_papers — Side-by-side paper comparison.
• index_dataset — Index a HuggingFace dataset card.
• search_datasets — Semantic dataset search with cross-encoder rerank.

About MCP

The Model Context Protocol (MCP) is an open protocol Anthropic introduced for connecting LLM-based agents to external tools and data sources. Providers publish MCP servers that expose their API surface as structured, discoverable tools — an MCP-compatible client (Claude Desktop, Cursor, Cline, Continue, etc.) can connect to the server and call its tools without any per-provider integration code.

Browse every MCP server on the APIs.io network or compare with the broader Agent Skill surfaces of the same providers.