Typed context & control research
TCCP
in-developmentStudy how source-bound context becomes a prompt, interface, or working artifact while keeping purpose, uncertainty, and the accountable next decision visible.
Research
Technical projects and research programs, explained with their current scope and limits. Start with a reader-friendly map, then follow the sources and methods that interest you.
These cards are orientation points. A program label describes the public-facing area; follow the linked project surface for implementation status and evidence.
Typed context & control research
Study how source-bound context becomes a prompt, interface, or working artifact while keeping purpose, uncertainty, and the accountable next decision visible.
Knowledge representation profile
Explore human-readable semantic document architecture with provenance, uncertainty, and review state before claiming a runtime or query service.
Fleet evaluation research
Compare method, role, and state-transition behavior with framework-neutral fixtures while retaining the scope and limits of each evaluation.
software
A local-first workbench for agent-supported research, source-linked knowledge, governance, and return. The Workbench page explains its current scope and boundaries.
specification
A notation for describing messages and bounded effects in human-agent work. Its grammar and validators make a description inspectable; they do not grant permission or prove the described workflow worked.
standard
A candidate framework for governed information, representation, symbolic expression, and execution in AI-mediated knowledge work. Its accepted architectural proposal remains distinct from standard adoption and implementation.
design-system
A public-facing design system exploring how layered, navigable surfaces can support different readers and return paths.
These are distinct lines of method work. The links lead to public introductions or proposals, not complete specifications or evidence of effectiveness.