Lilith CLI
Terminal IDE and AI-agent operator surface with local memory, inspectable tools and explicit delegation.
- Public proof
- Public · 4.6.0
- Limits of this case
- public
Project Intelligence 3
Nine projects with a stated purpose, public evidence and explicit limits. Open-source tools, private architecture and experiments are distinguished on each page.
Selected case studies
The strongest pages show the work as a bounded system: what it solves, what stays private and what evidence is public.
Terminal IDE and AI-agent operator surface with local memory, inspectable tools and explicit delegation.
Private nine-realm local ecosystem with a deliberately non-operational public architecture model.
Private desktop creative-operations system for projects, candidates, review, technical QA and publication gates.
Terminal IDE and AI-agent operator surface with local memory, inspectable tools and explicit delegation.
Public engineering fork for concrete agent compatibility work and upstream contribution.
Private nine-realm local ecosystem with a deliberately non-operational public architecture model.
Private creative-production framework with durable evidence and separate generation, review, promotion and release authority.
Private psychological-horror game experiment using camera language, controlled ambiguity and environmental discomfort as systems.
Private desktop creative-operations system for projects, candidates, review, technical QA and publication gates.
Private procedural-horror RPG experiment where deterministic state and local AI support a world that reacts to what the player leaves unsaid.
Private local assistant architecture built around explicit IPC, approvals and bounded system tools.
Private model-orchestration layer for routing work across local and external model providers under evidence and policy constraints.