별도 채점 하네스를 만들지 않고도 산출물을 재검산 가능하게 비교할 수 있도록 고정 100점 rubric, opaque 평가, 직접 evidence와 단순 총점 규칙을 마지막 작업으로 추가한다. |
||
|---|---|---|
| .. | ||
| advisor-context-hook-second-wave.md | ||
| inference-api-surface-execution-lifecycle-refactor.md | ||
| knowledge-tool-validation-optimization.md | ||
| llm-judged-missing-tool-call-retry-gate.md | ||
| long-term-memory-rag-second-wave.md | ||
| openai-compatible-hybrid-request-execution-routing.md | ||
| openai-compatible-incomplete-tool-call-syntax-gate.md | ||
| openai-compatible-output-validation-filters.md | ||
| openai-compatible-runtime-output-integrity-filter.md | ||
| rag-local-routing-model-operations.md | ||
| request-context-assembly-optimization.md | ||
| thin-agent-model-comparison-benchmark.md | ||
| tool-call-validator-model-gate-review.md | ||