A governance practice, packaged.
Thirty-eight skills and forty-five governance tools that run a data governance engagement end to end, with an audit trail a regulator could read.
The problem it solves
Data governance is the discipline most often bought and least often felt. Programs produce frameworks, committees, and slideware, then stall because nothing connects the policy layer to the day the data actually moves.
The fix is not more framework. It is an assembly line: a repeatable sequence of phases where each one produces a specific artifact, every decision is recorded with its evidence, and the whole thing can be shown to an auditor without a scramble.
What I built
Thirty-eight skills organized around the six phases of the assembly line, plus thirteen vertical packs carrying the regulatory reality of specific industries, and forty-five governance tools underneath.
Six phases over one governed substrate. The substrate is what makes the phases auditable rather than merely sequential.
The phases
Discovery runs from pre-meeting preparation through conduct and synthesis. The blueprint phase sets the operating model. Architecture covers platform and catalog. Build handles integration and automation. Operate covers KPIs and stewardship. And the final phase is AI enablement, because the point of governance in 2026 is that AI can be trusted with the data.
The substrate
This is the part that separates it from a content library. Underneath every phase runs a governed layer: a cryptographically signed decision log with tamper detection and supersession chains, a risk-tier rubric and classifier, a business-logic-rule catalog with a verifier, a compliance citation detector, engagement health snapshots across seven dimensions, and an audit pack generator that assembles the whole record on demand.
The citation work is a good example of the standard. The audit pack does not dump every regulatory mention it can find. It scopes them, so a healthcare engagement's pack shows the in-scope citations and suppresses the ones that matched but do not apply, with the counts shown so the reviewer can see the filtering happened.
The verification layer
A fifty-two test regression suite, continuous integration with a nightly schedule, manifest and body drift detection, schema stability verification, and a cross-suite alignment verifier that checks this system against its sibling. Four engagement workspaces ship with it, including a fully populated reference engagement so a new consultant can see what good looks like before touching a client.
That reference engagement is explicitly tagged as fixture content in every file, not client material. That labeling is deliberate. A realistic fixture that could be mistaken for real client work is a liability.
What it gives an organization
- A governance program that produces artifacts on a schedule rather than committees.
- An audit trail that exists as a by-product of delivery rather than a project of its own.
- Industry regulatory context carried in the tooling rather than in one specialist's memory.
- A path from governance to AI enablement, which is the reason most clients are asking in the first place.
General availability at v2.8.0, with the regression suite green. Deferred items are documented with their blocking reasons rather than quietly dropped.