wicked-estate

wicked-estate

Turn a repo — and its surrounding infrastructure/mainframe estate — into one queryable graph that LLM agents can actually use. Symbols, calls, imports, types, refs, and cross-domain estate links: definitions, who-calls-X, blast-radius, scoped context. Local-first, single static binary, tree-sitter + SQLite.

Status: working engine — cargo test --workspace is 617 passing, 0 failed, 0 ignored; 0 build warnings; clippy -D warnings clean. Greenfield, pre-1.0. See FEATURES.md for the exhaustive, honestly-tagged capability inventory (✅ built / 🟡 partial / 🟦 designed-not-built).

Why

LLM coding agents waste turns grepping and re-reading files. wicked-estate gives them a precise, ranked, bounded answer instead — "who calls this?", "what breaks if I change it?", "give me just the context for this symbol" — across 91 wired languages plus a mainframe/IaC estate layer (COBOL, JCL, RACF, IMS, MQ, Terraform, CloudFormation, …) that almost nothing else unifies into one graph. Every edge carries {confidence, provenance, resolved_by}; heuristics are never presented as facts.

Install

# From crates.io (once published):
cargo install wicked-estate

# Or straight from this repo (works today):
cargo install --git https://github.com/mikeparcewski/wicked-estate wicked-estate

# Or build from source:
git clone https://github.com/mikeparcewski/wicked-estate
cd wicked-estate && cargo build --release   # binary at target/release/wicked-estate

Real semantic search is opt-in (keeps the default build dependency-free + offline):

cargo install wicked-estate --features model2vec   # static embeddings, light, no ONNX
cargo install wicked-estate --features fastembed   # contextual ONNX/BGE, highest quality

Quickstart

# Index a repo into a local graph
wicked-estate index ./my-project --db graph.db

# Who/what depends on a symbol (blast radius = transitive dependents)
wicked-estate blast-radius MyClass --db graph.db

# Find a symbol + print its source
wicked-estate query handleRequest --db graph.db
wicked-estate source handleRequest --db graph.db

# Most important symbols (PageRank)
wicked-estate rank --db graph.db

# Semantic search (after: wicked-estate index … --embeddings, built with a semantic feature)
wicked-estate semantic "login security check" --db graph.db

# Stats (incl. git provenance), keep it watching, or stream changes
wicked-estate stats --db graph.db
wicked-estate watch ./my-project --db graph.db

What it does (highlights — full list in FEATURES.md)

• Code graph — 91 wired languages (tree-sitter), 100 in the manifest; symbols, calls, imports, heritage. Languages are data (a manifest row + a .scm query) — adding one is zero core change.
• Precise blast-radius — bounded reverse-reachability over all dependency edge kinds (not just calls), so it never silently under-reports.
• Layered resolution — name / scoped / import-map → SCIP (precise) → on-demand LSP. Two-phase EXTRACT → RESOLVE; resolution is swappable without re-parsing.
• Estate mapping — IaC (Terraform/CloudFormation/K8s) + mainframe (RACF security, IMS data, MQ messaging) as just-more-languages, joined cross-domain (e.g. the RACF profile that protects the dataset a JCL step uses — in one query). Drift = graph diff iac vs live.
• Hybrid + semantic retrieval — graph + FTS5 core, embeddings an optional sidecar fused via RRF. Three embedder tiers: lexical (default) → model2vec (static, light) → fastembed (ONNX/BGE).
• MCP server — exposes the retrieval tools to agents over JSON-RPC, following a strict runtime-behavior contract (cap output, report staleness, never error-early, label confidence).
• Requirement ↔ code linking — annotate nodes with description / requirement / requirement_validated and query by requirement.
• Git-aware + incremental — per-file git sha, incremental re-index, watch mode, a read-only edge-history log, and a resumable change-log (subscribe).
• Multi-repo — federated cross-graph search + blast-radius across many repo graphs.

Architecture

The whole engine programs against five traits (wicked-estate-core); everything else is a swappable impl behind a seam: