LLM-compiled knowledge bases for any AI agent. Parallel multi-agent research, thesis-driven investigation, source ingestion, wiki compilation, querying, and artifact generation.

Let AI agents investigate and operate infrastructure through declared actions, bounded by policy and host-side checks. Approvals when required, with an audit trail. By Protectorate.

Forcing-function skills for AI agents — validated to improve behavior, not just change it.

Claude/Codex agent skills distilled from @aleabitoreddit (Serenity)'s full public archive — track her, analyze like her, anticipate her next focus. Bilingual 中文/English.