Plan: Native CLI and tooling (native/cli) :: Vault

Problem statement

An optional D CLI exists for experiments (dub project). The primary product is TypeScript/Electron. Without a clear plan, the CLI risks duplicating logic, drifting from OpenAPI contracts, or confusing contributors about where features belong.

Current state (as-is)

D project under native/cli with documentation skeleton.
No required coupling to desktop release artifacts in CI (verify against current ci.yml when implementing phases).

Target state (to-be)

Explicit charter: what the CLI is for (e.g. batch validation, fixture generation, perf benchmarks) vs what must live in Electron.
Release story if binaries are distributed (versioning, platform matrix) — or explicit “source-only tool.”

Phase 0 — Charter ADR

Write 1-page ADR: keep / shrink / archive CLI; success metrics (e.g. “CI runs dub test on PRs”).

Acceptance: ADR merged; README links to it from native/cli README if present.

Phase 1 — Contract alignment

Commands that validate bundles against JSON Schema from disk.
Optional codegen helpers that stay in sync with OpenAPI (or explicitly defer to packages/sdk).

Phase 2 — CI integration (optional)

Add job matrix entry only if maintenance cost is acceptable (D compiler cache, dub fetch time).

Dependencies

data model plan if CLI operates on vault files.

Risks

Contributor friction from multi-language repo — keep D surface minimal or document Dockerized workflow.