jrswab/axe
A ligthweight cli for running single-purpose AI agents. Define focused agents in TOML, trigger them from anywhere; pipes, git hooks, cron, or the terminal.
What's novel
A ligthweight cli for running single-purpose AI agents. Define focused agents in TOML, trigger them from anywhere; pipes, git hooks, cron, or the terminal.
Code Analysis
6 files read · 2 roundsA sophisticated Go-based agent orchestration framework that dynamically chains sub-agents via LLMs, manages persistent Markdown memory, and integrates with MCP servers for tool execution.
Strengths
Exceptional modularity with a clear separation between core engine, agents, tools, and memory layers. The architecture elegantly handles dynamic tool injection (including MCP) and robust error propagation using a Result pattern. Memory persistence via atomic file writes is a practical and reliable design choice.
Weaknesses
Test coverage appears moderate; while functional tests exist, they may not exhaustively cover all edge cases in the LLM interaction loops or complex memory trimming scenarios. Some complexity in the sub-agent execution flow could benefit from further abstraction if tool sets grow significantly.
Score Breakdown
Signal breakdown
Innovation
Craft
Traction
Scope
Evidence
Commits
87
Contributors
3
Files
204
Active weeks
3
Repository
Language
Go
Stars
600
Forks
12
License
Apache-2.0