Axme Code
Provides Claude Code agents with persistent project memory, architectural decisions, and critical safety guardrails across all sessions.
AXME Code solves the problem of Claude Code agents forgetting project context between sessions by offering persistent memory, architectural decisions, and robust safety guardrails. It transparently ensures your AI coding agent always remembers your project's stack, decisions, and history, eliminating the need to repeatedly explain architecture. The tool also implements hard enforcement hooks to prevent dangerous commands, freeing developers to focus on building while ensuring session continuity and project integrity.
主な機能
010 GitHub stars
02Comprehensive suite of 19 MCP tools for agent interaction and management
03Automatic Knowledge Extraction from session transcripts for continuous learning
04Persistent Knowledge Base for continuous context (oracle, decisions, memory, backlog, handoff)
05Reliable Safety Guardrails that block dangerous commands before execution
06Support for Multi-Repo Workspaces with merged and scoped context
ユースケース
01Maintaining consistent project context for AI coding agents across multiple sessions
02Enforcing coding standards, architectural decisions, and deployment rules in AI-driven development
03Preventing AI agents from executing dangerous or unauthorized commands like `git push --force`
04Automating the collection of project knowledge, decisions, and lessons learned from agent interactions