test(cli): cover Agent Skills expectations criteria

[christso]

Summary

• Adds regression coverage for Agent Skills evals.json cases that use expectations without assertions.
• Verifies the read adapter emits agent-skills-criteria llm-rubric entries with stable expectation-N IDs and required: true.

Notes

Latest origin/main already moved Agent Skills evals.json support out of the removed core loader and into the CLI read adapter, and it already parses expectations. This PR preserves that behavior with focused expectations-only coverage.

Verification

• bun test apps/cli/test/commands/convert/convert-evals-json.test.ts - pass
• bun run lint - pass
• git diff --check - pass
• bun test packages/core/test/evaluation/orchestrator.test.ts -t "applies exponential backoff between retries" - pass after full-suite timing failure
• bun run test - one timing-sensitive failure in packages/core/test/evaluation/orchestrator.test.ts (applies exponential backoff between retries); reran that exact test successfully

Review

Subagent review reported no findings. Residual gap noted: this new case exercises the read-adapter object path; the existing basic conversion test already covers serialized YAML for expectations when assertions are also present.

Closes av-p60u.

[cloudflare-workers-and-pages]

Deploying agentv with    Cloudflare Pages

Latest commit:	c7d7048
Status:	✅  Deploy successful!
Preview URL:	https://88efeb01.agentv.pages.dev
Branch Preview URL:	https://task-av-p60u-agent-skills-ex.agentv.pages.dev

View logs