Adopt pyright: basic mode in CI, then targeted strict

[pbean]

Finding F-4 of the "Brownfield CLI/TUI Refactor Assessment" (local report, gitignored under _bmad-output/implementation-artifacts/).

Evidence

No type checker anywhere: no pyright/mypy config in pyproject.toml, .trunk/, or CI. Annotations are pervasive but unverified, so annotation drift is invisible and the refactor program's code moves lose a cheap safety net.

Recommendation — two stages

Stage 1 (basic mode):

1. Spike: run pyright in basic mode over src/bmad_loop (exclude src/bmad_loop/data/) and count errors.
2. If the burn-down is >~200 errors, STOP and report the count here with a proposed per-module rollout instead of landing the gate.
3. Otherwise: fix cheap errors, add targeted per-file suppressions for the rest (each with a one-line reason), commit pyright config (pyproject [tool.pyright] or pyrightconfig.json), and add a CI job mirroring the existing lint job structure in .github/workflows/ci.yml.

Stage 2 (targeted strict): enable per-module strict checking for the pure modules only (model.py, statemachine.py, policy.py, documents.py, machine.py, checks.py, sanitize.py); fix or precisely suppress what surfaces.

No runtime behavior changes in either stage.

Ordering

Sessions 8 and 12 of the refactor program. Stage 1 should land before the Engine collaborator extractions (it fences those moves); Stage 2 requires Stage 1 merged.

Traps

• Verify current pyright config options/versions against official docs before writing config — don't pin from memory.
• The burn-down size is an explicitly deferred decision until the spike runs.

[coderabbitai]

🔗 Related PRs

#177 - fix(engine): strip the stale Auto Run Result before every review launch (#160) [merged]

---

📝 Issue Planner

_{Check the box below or use the @coderabbitai plan command to generate an implementation plan and prompts that you can use with your favorite coding assistant.}

• Create Plan

---

🧪 Issue enrichment is currently in open beta.

You can configure auto-planning by selecting labels in the issue_enrichment configuration.

To disable automatic issue enrichment, add the following to your .coderabbit.yaml:

issue_enrichment:
  auto_enrich:
    enabled: false

💬 Have feedback or questions? Drop into our discord!

[pbean]

Stage 2 (per-module strict on the seven pure modules — model, statemachine, policy, documents, machine, checks, sanitize) is delivered in #268, completing this issue. Stage 1 (basic-mode gate) shipped in #264, which auto-closed the issue; recording the Stage-2 close-out here so the trail is complete. uvx pyright@1.1.411 is at 0 errors with the strict list in place; no runtime changes.