Closes the first wave of every trust-system arc (Arcs 1-7) plus the two
wave-2 follow-ups landed the same day. Nine feature PRs since v1.10.0.

- Arc 1 (#170): rocky lineage --downstream, rocky branch, rocky run --branch, rocky replay
- Arc 2 (#171): per-run cost attribution, [budget] block, budget_breach hook
- Arc 3 (#172): three-state CircuitBreaker, adapter consolidation
- Arc 4 (#173): rocky trace Gantt + feature-gated OTLP metrics export
- Arc 5 (#174): schema-grounded rocky ai prompt + project-aware validator
- Arc 6 wave 1 (#184): --target-dialect P001 portability lint (12 constructs)
- Arc 7 wave 1 (#185): blast-radius P002 SELECT * lint (semantic-graph aware)
- Arc 6 wave 2 (#186): [portability] config block + per-model rocky-allow pragma
- Arc 7 wave 2 wave-1 (#187): --with-seed source-schema inference

Plus #169 fix: install scripts pick latest engine version by semver.

Version bump: 20 Cargo.toml files (all workspace members except
rocky-bigquery, which tracks its own version).

Wave 2/3 work for every arc remains in the deferred backlog — see
the changelog Deferred section for the full carry-forward.

…202)

Adds a new `rocky cost <target>` CLI verb that reads `RunRecord` from
the embedded state store and rolls per-model cost attribution up from
persisted `bytes_scanned`, `bytes_written`, and `duration_ms` values.
Re-uses `rocky_core::cost::compute_observed_cost_usd` — the same
formula `RunOutput::populate_cost_summary` applies at the end of a
live run — so the historical surface stays consistent with the
first-wave per-run summary (PR #171).

The command loads `rocky.toml` to resolve the billed-warehouse type;
when the config can't be read the output degrades gracefully to
`adapter_type: null` / `cost_usd: null`, and durations/bytes are still
emitted from the stored record. BigQuery is a quiet upgrade: because
`ModelExecution.bytes_scanned` is persisted, the historical command
returns a real cost figure for BQ runs even though the live run path
still reports `null` for BQ (adapter bytes-scanned plumbing is a
later wave).

- New `CostOutput` + `PerModelCostHistorical` in `output.rs` (deriving
  `JsonSchema`), registered in `export_schemas.rs::schemas()`.
- New `commands/cost.rs` with `run_cost` entry point and 12 unit tests
  (rollup math, empty-adapter degrade, BQ bytes path, model filter,
  by-id and latest resolution, adapter-default precedence).
- Clap subcommand `Cost { target, model }` in `rocky/src/main.rs`.
- Regenerated `schemas/cost.schema.json`, dagster Pydantic
  (`types_generated/cost_schema.py`), and vscode TypeScript
  (`types/generated/cost.ts`) via `just codegen`.

Scope notes:
- Does NOT extend `ReplayOutput`, `RunOutput`, or any existing Output
  struct — the new surface is isolated to `CostOutput`.
- Does NOT introduce per-model `[budget]` blocks or the PR
  cost-projection Action — those are later waves.
- The write side (`StateStore::record_run` being called from the live
  `rocky run` path) is a pre-existing gap shared with `rocky replay`,
  `rocky trace`, and `rocky history` — out of scope for this PR.
- Dagster resource-method wiring is deferred; the Pydantic type is
  reachable today via `types_generated.cost_schema`.