update profiles

Removed slugs to meet 5 slug limit

Fixed all clippy issues

GOAL: Address all clippy issues with strictest settings

1. Complex type warnings in async-mcp crate:
   - completable/mod.rs
   - protocol.rs
   - registry.rs
   - server/resource/mod.rs
   - server/roots/mod.rs
   - server/sampling/mod.rs
   - server/tool/mod.rs
2. Needless borrows and redundant conversions
3. Unnecessary lifetimes in multiple files
4. Other clippy warnings (map_or simplification, manual stripping, etc.)

GOAL: Integrate actix-web-lab SSE transport into the async-mcp crate

IMPLEMENTATION:
1. Added required dependencies to Cargo.toml:
   - actix-web-lab = "0.20"
   - bytes = "1.5"
   - bytestring = "1.3"
   - futures-core = "0.3"
   - futures-util = "0.3"
   - pin-project-lite = "0.2"

2. Updated SSE transport implementation:
   - Replaced custom SSE implementation with actix-web-lab's SSE
   - Added new methods for handling SSE events and messages
   - Implemented proper error handling and type conversions
   - Added test cases for SSE transport functionality

3. Updated HTTP server integration:
   - Modified SSE handler to use new transport implementation
   - Fixed imports and module organization
   - Improved error handling and logging
   - Added proper endpoint information handling

4. Fixed module organization:
   - Updated transport module exports
   - Fixed import paths in http_transport.rs
   - Ensured proper visibility of types

Key improvements:
- More robust SSE implementation using battle-tested actix-web-lab
- Better error handling and type safety
- Improved code organization and modularity
- Maintained backward compatibility with existing interfaces

GOAL: Improve async-mcp benchmark implementation to provide more accurate and stable measurements

IMPLEMENTATION:
1. Organized benchmarks into logical groups:
   - completions: string and fixed completion benchmarks
   - notifications: notification sending benchmarks
   - prompt: prompt builder benchmarks
   - server: connection and capabilities benchmarks

2. Improved benchmark methodology:
   - Added proper warm-up time (1 second)
   - Pre-allocated test data to avoid allocation overhead
   - Used BatchSize for proper setup/teardown
   - Added throughput measurements
   - Implemented zero-overhead mocks

3. Added benchmark helper functions:
   - bench_async: consistent async benchmark runner
   - Grouped related benchmarks for better comparison

4. Removed artificial delays:
   - Eliminated sleep in notification sender
   - Simplified mock implementations

Implemented complete mcp spec + openai/ollama bridge