v1.0.0

What's changed

• Run the skill evals on an API key, keeping the WIF path one variable away
• Let users and agents report plugin bugs and gaps as GitHub issues
• Stabilize the not-invented-id eval and correct the temperature-0 determinism claim
• Run the code review in a fresh agent from a shipped prompt, with pluggable project rules
• Give the review report a fixed, scannable layout

Full changelog: v0.1.6...v1.0.0