ai-nexus 1.4.8 → 1.4.9
This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
- package/README.ko.md +1 -1
- package/README.md +1 -1
- package/package.json +1 -1
package/README.ko.md
CHANGED
|
@@ -21,7 +21,7 @@ npx ai-nexus install
|
|
|
21
21
|
|
|
22
22
|
AI 코딩 도구마다 룰 형식이 다릅니다 — `.claude/rules/*.md`, `.cursor/rules/*.mdc`, `.codex/AGENTS.md`. 같은 룰을 여러 곳에 관리하다 보면 결국 서로 달라집니다. 게다가 매 프롬프트마다 모든 룰이 로드되어 불필요한 토큰이 낭비됩니다.
|
|
23
23
|
|
|
24
|
-
[ETH Zurich의 최근 연구](https://arxiv.org/pdf/2602.11988) (
|
|
24
|
+
[ETH Zurich의 최근 연구](https://arxiv.org/pdf/2602.11988) (12개 레포, 5,694개 PR 대상)가 이를 확인합니다: **모든 룰을 한꺼번에 로드하면 성능이 ~3% 하락하고 비용은 20% 이상 증가합니다.** 사람이 직접 작성한 컨텍스트 파일도 30줄 이하일 때만 4% 개선 효과가 있었습니다. 결론: 프롬프트마다 관련 룰만 로드해야 합니다.
|
|
25
25
|
|
|
26
26
|
## 해결
|
|
27
27
|
|
package/README.md
CHANGED
|
@@ -21,7 +21,7 @@ npx ai-nexus install
|
|
|
21
21
|
|
|
22
22
|
Every AI coding tool has its own rule format — `.claude/rules/*.md`, `.cursor/rules/*.mdc`, `.codex/AGENTS.md`. You end up maintaining the same rules in multiple places, and they inevitably drift apart. On top of that, every prompt loads all your rules, wasting tokens on irrelevant context.
|
|
23
23
|
|
|
24
|
-
A [recent study from ETH Zurich](https://arxiv.org/pdf/2602.11988) (
|
|
24
|
+
A [recent study from ETH Zurich](https://arxiv.org/pdf/2602.11988) (12 repos, 5,694 PRs) confirms this: **loading all rules at once hurts performance by ~3% and increases cost by 20%+.** Even hand-written context files only helped by 4% — and only when kept under 30 lines. The takeaway: less is more, and only relevant rules should be loaded per prompt.
|
|
25
25
|
|
|
26
26
|
## The Solution
|
|
27
27
|
|
package/package.json
CHANGED