RoslynLens

Comparison with alternatives

How RoslynLens compares to other Roslyn-based MCP servers.

Feature matrix

Feature RoslynLens SharpLensMcp RoslynMCP SharpToolsMCP CodeAnalysisMCP
Tools 34 58 5 ~20 3
Anti-pattern detectors 19 0 0 0 0
Roslynator analyzers (opt-in) Yes (500+ rules) No No No No
Complexity metrics Yes Yes Yes Yes No
Data flow analysis Yes Yes No No No
Duplicate detection Yes No No No No
Compound tools Yes Yes No No No
Batch operations Yes Yes No No No
Convention validation Yes No No No No
Circular dependency detection Yes No No No No
Test coverage mapping Yes No No No No
Glob pattern search Yes Yes Yes No No
Fuzzy FQN resolution Yes No No Yes No
External source resolution Yes No No Yes No
Refactoring tools No Yes (13) No Yes (6) No
Auto solution discovery Yes (BFS + multi-solution) No No No No
Global dotnet tool Yes Yes No No No
File watcher Yes No No No No
MCP tool annotations Yes No No No No
Reproducible token benchmark Yes No No No No
Docker image Yes No No No No
LRU compilation cache Yes (configurable) Yes Yes No No
Token-optimized responses Yes (30-150 tokens) Relative paths No Stripped indentation No
.NET version 10 8 8 8 9
License Apache-2.0 MIT MIT MIT MIT

Key differentiators

RoslynLens strengths

  1. 19 anti-pattern detectors — No other MCP server provides automated code quality analysis. Covers async, security, EF Core, logging, and domain-specific patterns.

  2. Convention validation — Check naming, security, and dependency rules against your team's standards.

  3. Auto solution discovery — BFS traversal finds your .sln/.slnx automatically. No manual path configuration needed.

  4. Circular dependency detection — At both project and type level.

  5. Test coverage mapping — Heuristic mapping between production types and test classes.

  6. Token efficiency — Responses are 30-150 tokens. Minimal property names, no unnecessary data. A reproducible, offline benchmark (docs/BENCHMARKS.md) measures a 77% pooled / 73% median token reduction versus reading source, self-hosted on this repository.

  7. Optional Roslynator analyzers — get_diagnostics can run the bundled Roslynator ruleset (500+ rules) on demand, on top of the curated detectors.

What others do better

  1. SharpLensMcp — 13 refactoring tools (rename, extract method, etc.). RoslynLens is read-only by design.

  2. SharpToolsMCP — Git integration with automatic undo branches. Good for AI-driven refactoring workflows.

Architecture comparison

Aspect RoslynLens SharpLensMcp SharpToolsMCP
Transport stdio stdio stdio + SSE
Discovery WithToolsFromAssembly() Manual registration Manual registration
Workspace MSBuildWorkspace + LRU MSBuildWorkspace MSBuildWorkspace
File sync File watcher (auto) Manual sync_documents() Git-based
Responses Typed records (JSON) Structured JSON Diffs + errors
Config Environment variables Environment variables CLI flags