auto_lift
One-call autonomous lifting via cheap LLM API (Haiku, GPT-4o-mini, OpenRouter, Gemini)
How to use it
auto_lift is exposed by the Rpg Encoder MCP server. Add the server to your MCP client (Claude Desktop, Cursor, Windsurf and others), and the auto_lift tool becomes available to the model automatically. See the full listing for setup details and every tool this server provides.
Install Rpg Encoder
claude mcp add rpg -- npx -y -p rpg-encoder rpg-mcp-serverOther tools in Rpg Encoder (26)
Code health: coupling, instability, god objects, clone detection
Index the codebase (run once, instant)
Get domain discovery + hierarchy assignment prompts
Single-call search + fetch + explore with token budget
Find circular dependencies and architectural cycles
Traverse dependency graph (upstream, downstream, or both)
Get entity metadata, source code, dependencies, and hierarchy context
Aggregate file-level features, rebuild hierarchy metadata
K-shortest dependency paths between two entities
Get entity source code for your agent to analyze
Get file-level entity features for holistic synthesis
Get entities needing semantic routing (drifted or newly lifted)
BFS reachability analysis — "what depends on X?"
Dashboard — coverage, per-area progress, NEXT STEP
Change planning — find relevant entities, modification order, blast radius
Dependency-safe reconstruction execution plan
Reload graph from disk after external changes
Graph statistics, hierarchy overview, per-area lifting coverage
Search entities by intent or keywords (hybrid embedding + lexical scoring)
Whole-repo semantic understanding in one call (~25K tokens for 1000 entities)
Extract minimal connecting subgraph between entities
Submit holistic file-level summaries
Apply hierarchy assignments to the graph
Submit the agent's semantic features back to the graph
Submit routing decisions (hierarchy path or "keep")
Incremental update from git changes