docs: update scoring methodology

[brth31]

• Hallucination/Security no longer default to 100 with zero graders on CLI-only evals (and Correctness no longer defaults to 0) — both become N/A and the remaining weights renormalize.
• Efficiency gets a command-precision variant for cli-only evals: measures probe-and-retry against the declared target operation instead of file-based waste that can't fire without files.
• Weight changes now require a before/after sensitivity check against stored results before merge.
• Baseline (zero-tool) demoted from the headline comparison — it conflates the generic lift of having any tool access with the specific lift of Auth0's tools, and likely overstates the published delta.
• Hallucination, Security, and Correctness given concrete, bounded criteria instead of open-ended labels.

[coderabbitai]

Important

• 🔍 Trigger review

This repository does not receive automatic reviews because it has fewer than 10 stars.

⚙️ Run configuration

Configuration used: defaults

Review profile: CHILL

Plan: Pro Plus

Run ID: 4abe2ffa-7b8a-441c-9b88-bf2e428b98bf

---

Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out.

❤️ Share

• X
• Mastodon
• Reddit
• LinkedIn

_{Comment @coderabbitai help to get the list of available commands.}