Add cost/token observability across loop tooling

[cobusgreyling]

Summary

Operationalizes cost/token awareness as a first-class part of loop engineering:

• Registry — per-pattern cost metadata (noop/report/action tokens, daily caps, early-exit flags)
• loop-init — scaffolds loop-budget.md, loop-run-log.md, and loop-budget skill on every pattern
• loop-audit — scores cost observability; caps L3 until budget + run log + LOOP.md budget exist
• loop-cost — new CLI for daily token spend estimates by pattern, cadence, and level
• Docs — Cost Profile sections in all 6 patterns; cost-aware picks in pattern-picker
• Dogfood — reference repo includes budget/run-log files and skills/loop-budget

Test plan

• npm run validate:registry
• npm run test:tools (loop-audit, loop-init, loop-cost)
• loop-audit on reference repo → L3 with full cost observability

[github-actions]

Loop Readiness Audit

Score: 100/100 (L3)

Strong loop readiness — good candidate for L3 with explicit gates.

No suggestions — looking good.

_{Posted by audit.yml · loop-audit docs}