What makes the workflows run.
Finding evidence, making it comparable, building systems and governing the work take different tools. Each has a defined job. Together they let a solution evolve without losing the reasoning and controls it depends on.
Explore how the tools work together
These pages describe the tools’ roles. Capabilities are at different stages of development; we confirm what is available when scoping the work.
The application framework that gives digital systems shared conventions for identity, data, provenance and permissions. Human developers can build on Frame directly; Forge provides the agent layer.
Builds and maintains digital systems on a chosen framework under human supervision. Reusable specifications and components carry what one build learns into the next.
Finds and checks public information across registers, reports and datasets. Its sector hubs connect that evidence into a useful public reference.
Maps documents, feeds and databases into a chosen schema, retaining the source trail and flagging interpretation that needs review.
Connects information across participating systems. Holders control discovery and access; DAAC brokers the connection without taking custody.
Builds and maintains websites, hub interfaces, reports and presentations. Reusable templates carry useful improvements from one build into the next.
Turns a client problem, research question or opportunity into a proposed solution and a brief for the work it needs.
Brings progress and operating signals into one view, so people can see what is being built, what is running and where attention is needed.
Compares specifications, code and the solutions built from them. It identifies disagreement and proposes a correction for the responsible owner.
Exercises a running system and compares its behaviour with requirements or a reference build. Its findings guide fixes and release decisions.
Dispatches work within defined permissions, records the result and brings decisions back for human review.
Operates the infrastructure behind the suite, including provisioning, patches, backups and health checks, through governed actions.
Checks that work is recorded in the right place and has someone responsible for reading and acting on it.
Routes a request for model capability to a provider through a shared contract, keeping provider choice separate from each tool’s implementation.
Provides a shared sign-in and session contract, so tools can authenticate people consistently.
Allows governed queries against a database through a read-only interface, without opening a second route for changing its data.
Compares capability needs with evidence from actual runs and changes in the model market, then proposes allocation changes for review.
Identifies recurring work that could become a reusable tool and defines how that tool should be built and governed.
Finds prior knowledge through text search and the relationships recorded between documents, so decisions can draw on work already done.
Brokers access to credentials through scoped grants. Workers request the authorised action without needing to handle the underlying secret.
Defines a governed route for outbound communication, separating actions that may run automatically from those requiring human release.
Routes user feedback to the work that can address it, distinguishing operational faults from requests for new capability.
Presents a decision and records the judgement against it, keeping the reason available to the people who act on it.
This site is a Mason build. What each build learns goes back into the templates the next site and hub front door fork from.