apex solar kiss is the builder practice. ASK builds governed domain architectures: systems that make governing intent, constraints, references, and standards explicit, and structure the authorized exercise of judgment to keep generative work explainable, auditable, and answerable in its domain.
A deterministic governance architecture brackets the generative execution span. The governing rules and decision rights stay explicit while generation varies inside the bound, and human judgment stays at the apex rather than merely in the loop. The underlying primitive is bounded generativity: define the grammar, permit bounded variance, make selection and governance explicit.
An AI agent can interpret a brief, devise a plan, select tools, and adjust its approach while working. Even so, the design problem is keeping that interpretation within the mandate: latitude to improve the work must not silently become authority to remove its obligations or redefine its purpose. Governance, in this sense, means the arrangements that establish who may decide what, under which authority and limits, how decisions and actions are checked, challenged, or corrected, and who remains answerable. The aim is useful latitude under a defined mandate, not manual approval of every move.
ASK began as a corporate lawyer specializing in banking and structured finance. That background informs how the practice specifies decision rights, delegated authority, and limits in agentic systems. Legal drafting, in turn, shapes how the practice articulates reusable concepts that help people and agents interpret tasks, examine unfamiliar cases, and coordinate a response. The practical grounding, however, comes from creative production and systems work: which reference governs, what may vary, who may accept a result, and what has to survive the handoff. The architecture connects those decisions to the tools carrying out the work; precise language does not, by itself, make the tools comply.
The practice is one thing; the projects are several. The current portfolio includes two domain applications and the builder machine used to build and govern them. All of it is open source.
domain applications
asset-pipeline-ASK
AP-ASK articulates and encodes a brand’s visual language as a clear, reusable system for image production. It draws on the brand’s imagery, references, and products, together with the accumulated knowledge of its creative team. It separates what must remain true from what may vary creatively, and makes explicit what each reference is meant to guide. It carries the brand’s direction from the brief through production, review, selection, and approval. Because AP-ASK is tool-independent, that direction can inform the work of photographers and art directors, direct AI image models, or be built into an Airtable-based workflow. The point is to not let the tool define the brand, but rather to keep the brand’s own creative direction intact as the tools change.
AP-ASK is a developing open-source reference architecture, not a finished production platform.
urban-observatory
Public-planning implementation intelligence. It reads public data to reconstruct how civic commitments move from decision into delivery, showing progress, delay, change, and uncertainty rather than collapsing implementation into a single verdict — and recording what the evidence cannot show as an observability limit.
the builder machine
The method, execution protocol, and design architecture used to build and govern the applications above. Each lives in a separate open-source repository.
method-ASK
The method: how to design systems that build systems, carried as doctrine, topology, and worked examples above any single project.
control-surface
The execution protocol: the shared workflow rules, review gates, and templates that govern how work gets done in a session, resolved locally into each project's operating rules.
design-system-ASK
The design architecture: the visual foundations shared across the ASK family — tokens, type, and the wordmark. Downstream surfaces use them without redefining them. A-S-K.studio vendors those foundations and the surface-shell pattern from design-system-ASK.
research artifacts
Structured surveys published as explorable instruments rather than as prose. Each states what it settles, what it does not, and on what evidence.
consciousness + free will
A mind map over a typed relation atlas: 548 primary objects, 482 evidence owners, and the governed relations between them. It surveys what the obtained literature does and does not settle about consciousness and free will, and where those questions touch the source-of-intent architecture. A structured survey, not doctrine — no question is wholly answered, and the six held questions are shown with ASK's rulings labelled as rulings.