Defines a modular browser-oriented architecture in which narrow model or adapter roles are selected from an installed library while only a bounded active set participates in a request.
Architecture guide: this topic defines a modular tiny-model planning contract. It does not claim that model artifacts exist, are compatible, or execute on this WordPress site.
System shape
A request enters an orchestration layer that proposes one or more capability roles. The controller resolves dependencies and compatibility, estimates local cost, activates the smallest acceptable set, obtains results, applies verification or aggregation, and records the action and evidence.
Why modular
Narrow specialists can reduce irrelevant capacity and make failure boundaries more inspectable. They also introduce routing errors, multiple licenses, version skew, duplicated tokenization, handoff loss, and more complex observability. Modularity is therefore a trade, not an automatic quality improvement.
Core records
- Installed registry and artifact provenance.
- Active-set state and residency.
- Dependency and compatibility graph.
- Resource ledger and candidate-action scores.
- Per-role quality, latency, and failure evidence.
- Rollback and retirement history.
Scope
This starter page defines the questions, boundaries, evidence, and failure modes that should be recorded before a capability is presented as supported.
Engineering considerations
- Identify the source, version, target environment, and owner.
- Separate observed values from estimates and externally reported values.
- Record trade-offs, unsupported cases, and fallback behavior.
- Link performance statements to a compatible benchmark methodology.
Verification questions
- What exact artifact, revision, backend, and environment were reviewed?
- Which assumptions could change the result?
- Which data should be retained so another engineer can reproduce the conclusion?