Hi! Suggesting three production-tested tools for Claude Code:
openclaw-self-healing ★32 — https://github.com/Ramsbaby/openclaw-self-healing
4-tier autonomous crash recovery (preflight → restart → watchdog → AI repair → alert). 64% auto-resolved in production. LLM-agnostic (Claude/GPT-4/Gemini/Ollama). Prometheus metrics, Docker, one-line install.
openclaw-memorybox ★8 — https://github.com/Ramsbaby/openclaw-memorybox
Memory hygiene CLI. memorybox doctor ~/.claude finds bloat; memorybox split fixes it. -83% MEMORY.md size, prevents context overflow crashes. Zero deps, offline.
openclaw-self-evolving ★2 — https://github.com/Ramsbaby/openclaw-self-evolving
Weekly agent improvement. Scans Claude Code logs for retry loops + rule violations → proposes CLAUDE.md diffs. Zero API cost. Human approval required.
All MIT, all battle-tested 24/7 in production. They solve the three core problems of stable long-running agents: crash recovery, memory hygiene, behavior learning.
Hi! Suggesting three production-tested tools for Claude Code:
openclaw-self-healing ★32 — https://github.com/Ramsbaby/openclaw-self-healing
4-tier autonomous crash recovery (preflight → restart → watchdog → AI repair → alert). 64% auto-resolved in production. LLM-agnostic (Claude/GPT-4/Gemini/Ollama). Prometheus metrics, Docker, one-line install.
openclaw-memorybox ★8 — https://github.com/Ramsbaby/openclaw-memorybox
Memory hygiene CLI.
memorybox doctor ~/.claudefinds bloat;memorybox splitfixes it. -83% MEMORY.md size, prevents context overflow crashes. Zero deps, offline.openclaw-self-evolving ★2 — https://github.com/Ramsbaby/openclaw-self-evolving
Weekly agent improvement. Scans Claude Code logs for retry loops + rule violations → proposes CLAUDE.md diffs. Zero API cost. Human approval required.
All MIT, all battle-tested 24/7 in production. They solve the three core problems of stable long-running agents: crash recovery, memory hygiene, behavior learning.