feat(routing): calibrate model tier defaults

[Phelan164]

Summary

Update the default subagent routing and example inventory:

Tier	Model	Reasoning
easy	gpt-5.6-sol	medium
medium	gpt-6-astra	low
difficult	gpt-6-astra	high

Align the README and routing tests, including coverage for both Astra effort levels in the inventory. The existing medium reviewer minimum now selects Astra with low reasoning.

Validation

• python3 scripts/validate_repo.py
• Playground tests pass when the lab changed — not applicable; lab unchanged.
• Changed skills pass quick_validate.py — not applicable; no repository skill changed.
• Version-sensitive product claims link to official documentation

Full measurement and routing utility suite: 79 tests passed after integrating current main and correcting its stale Luna assertion. Playground: 2 tests passed. Wiki lint, measurement-template summary, Python compilation, and git diff --check passed.

Safety and scope

• Examples contain no credentials or destructive defaults
• Community practices are labeled as such
• The change is limited to the described scope

These remain customizable starting preferences. Verify model availability in the target runtime. This PR updates repository examples; existing private installations require a separate configuration update. Rollback: revert this commit.