Spec-kit extensions: persistence, guardrails, health states

[veerabhadra-ponna]

Really liking the new /clarify and /analyze commands; they already cover a lot (ambiguities + consistency checks). A few extra ideas that might make spec-kit even stronger:

• Context snapshots: preserve LM context in a lightweight artifact so sessions can be resumed or switched between models without losing the thread.
• Validation hooks: enforce guardrails from the constitution or catch contradictions automatically before agents run with the plan.
• Task/model routing metadata: mark which tasks go to lightweight vs heavyweight models for cost + reproducibility.
• Artifact health states: flag specs as draft/reviewed/consistent so it’s clear when they’re ready to act on.

Feels like these would complement clarify/analyze by adding persistence, guardrails, and reproducibility. Curious what others think.