AgentHUD – Live TUI and daily digest for parallel Claude Code sessions
LLM-generated daily digests of Claude Code activity is a clever observability twist.
Pre-warmed Claude Code sessions in tmux. No startup wait.
Session pooling for AI CLIs — why didn't this exist before?
Claude Code users who switch between agent roles frequently
llm-primer keeps a pool of warm sessions in the background. When you need one, it's already past the startup protocol.
Happy to answer questions. The main motivation was that I'm running LLM wikis (llm-context-base) and switching context between different agent roles constantly. The startup latency was killing the flow.
LLM-generated daily digests of Claude Code activity is a clever observability twist.
Pre-flight checks catch blockers before bed, stop hooks prevent 3am failures.
Automated IC screen is clever, but founders won't trust AI scoring over real feedback.
Budget control with auto-kill saves money when agents go rogue on long runs.
PTY injection and stop-hook parsing bypass Anthropic's expensive programmatic credit pool.
Recovers wiped projects from Claude Code logs when git history won't help.