Make /spec-kitty.analyze a true pre-implementation readiness review

## Problem

`/spec-kitty.analyze` currently behaves like a lightweight cross-artifact consistency check over `spec.md`, `plan.md`, `tasks.md`, and charter. That is useful, but it is too narrow for the role it occupies in the workflow: the last optional quality gate after `/spec-kitty.tasks` and before `spec-kitty next` starts agent execution.

By that point the mission has more artifacts than the command reviews: `meta.json` intent fields, `research.md`, `data-model.md`, `contracts/`, `quickstart.md`, optional planning artifacts such as `occurrence_map.yaml`, `tasks.md`, and the individual `tasks/WPxx-*.md` work-package prompts. The current prompt can miss big-picture user-intent drift, planning-artifact contradictions, bad WP decomposition, missing WP readiness metadata, and overbuilt or poorly sequenced delivery slices.

## Product owner view

This command should become a **Pre-Implementation Readiness Review**. It should answer: "Are we about to send implementation agents into a plan that actually matches the user's goal, is internally coherent, and is decomposed into executable WPs?"

## Desired outcome

`/spec-kitty.analyze` should review the full planning corpus and produce a concise readiness verdict before implementation begins.

Minimum review dimensions:

1. Intent fidelity: compare `meta.json`/source intent/purpose fields against `spec.md`, `plan.md`, `tasks.md`, and WPs.
2. Specification quality: FR/NFR/user-story/edge-case completeness and ambiguity checks.
3. Planning artifact consistency: `plan.md`, `research.md`, `data-model.md`, `contracts/`, `quickstart.md`, decisions, and optional mission-specific artifacts.
4. Work-package readiness: every WP prompt exists, is well-sized, has sufficient context, DoD, risks, dependencies, ownership, branch strategy, and reviewer guidance.
5. Coverage and mapping: every requirement maps to WPs; every WP maps back to user value/requirements; all planning artifacts are represented in tasks.
6. Execution readiness: use or recommend `finalize-tasks --validate-only` to catch dependency cycles, missing requirement refs, ownership overlaps, and invalid frontmatter before `spec-kitty next`.
7. Product coherence: MVP slice, sequencing, overbuild risk, missing user-value slice, and whether the plan still solves the original problem.

## Acceptance criteria

- The analyze command prompt is reframed as a pre-implementation readiness review after `/spec-kitty.tasks` and before `spec-kitty next`.
- It explicitly loads and reasons over the full mission planning corpus, not only `spec.md`, `plan.md`, and `tasks.md`.
- It reviews individual `tasks/WPxx-*.md` prompt files, not just the task index.
- It emits a readiness verdict: ready, ready-with-caveats, or blocked.
- It produces actionable findings grouped by severity and by artifact.
- It tells the user whether to proceed to `spec-kitty next` or return to specify/plan/tasks.

## Notes

This issue follows a product review of the current command template. Current docs already recommend `/spec-kitty.analyze` after `/spec-kitty.tasks`, but the command behavior is not broad enough to be trusted as the last planning gate.