The visible feature was only the surface
Adding a skill to a catalog touches duplicated consumers and can import stale package/repository facts or mistake "publish" for permission to deploy.
The system also had to respect concrete constraints. Catalog schemas and consumers vary by target app. Repository/package/version/compatibility facts require primary-source verification. Local catalog editing does not authorize commit, push, npm publication, pull requests, or deployment.
My role covered agent workflow design and automation engineering, with responsibility for agent workflow, validation automation, and packaging.
I turned the constraints into boundaries
Source identity/evidence resolution; target-app schema and consumer tracing; Python entry validator plus repository-native format/type/test/build/sitemap gates.
Discover the live catalog schema instead of assuming one.
Update exactly one entry and trace all existing consumers.
Pair the heuristic validator with the app's own checks because JSON shape alone cannot prove presentation or routing.
The key decision: Discover the live catalog schema.
The implementation had to prove the model
I delivered the work across the full path: Defined evidence and authorization rules. Built schema/consumer discovery workflow. Implemented deterministic catalog-entry validation. Integrated README index, sitemap, type-check, test, and production-build acceptance gates.
The result is concrete: Agents can resolve a canonical skill identity. Add/update a valid entry. Keep indexes and sitemap aligned. Report local completion without implying external publication.
The senior engineering lesson was that solving the visible workflow is only half the job. The architecture must also make constraints, failure modes, evidence, and ownership explicit enough for the next change to remain safe.