Hmbown/aleph
Skill + MCP server to turn your agent into an RLM. Load context, iterate with search/code/think tools, converge on answers.
What's novel
Skill + MCP server to turn your agent into an RLM. Load context, iterate with search/code/think tools, converge on answers.
Code Analysis
5 files read · 3 roundsA sophisticated, provider-agnostic framework for orchestrating multi-agent systems with advanced context management, sub-query memoization, and file-based knowledge retrieval.
Strengths
Exceptional modularity with a clean provider abstraction layer that enables seamless backend swapping. The sub-query system demonstrates deep substance through robust context slicing, content format detection, and intelligent caching strategies that prevent redundant API calls.
Weaknesses
Lack of visible test files suggests limited automated validation of the complex orchestration logic. The absence of a main entry point file in the explored sections makes it unclear how users interact with the system without relying solely on documentation.
Score Breakdown
Signal breakdown
Innovation
Craft
Traction
Scope
Evidence
Commits
114
Contributors
2
Files
122
Active weeks
11
Repository
Language
Python
Stars
176
Forks
19
License
MIT