TechVerser on X: Claude Code 的上下文终于不用再被一堆预加载技能塞满了！推荐一个非常优雅的按需加载解决方案：Jev Skill Suggestion。
@realfxw@realfxwlisted 49m agoreviewed by JevClaude Code 的上下文终于不用再被一堆预加载技能塞满了！推荐一个非常优雅的按需加载解决方案：Jev Skill Suggestion。 以往在配置大量自定义技能（Skills）时，最头疼的就是 Context Window 膨胀问题。许多低频技能即便一次都用不上，也会在每次对话启动时常驻上下文，不仅白白消耗 Token 成本，还会稀释模型的注意力，甚至引入不必要的提示词干扰。 这个 Mod 的核心逻辑是将技能路由与主体执行彻底解耦（Just-in-Time 动态注入）： 常态零占用：将技能标记为仅用户可调用，默认状态下完全排除在上下文窗口之外，保持 Prompt 绝对干净。 轻量前置路由：用户输入指令后，Mod 会将技能列表交由 TypeSafe AI 的分类器 Jev（或兼容的 Vercel AI Gateway），快速匹配与当前任务最契合的技能。 精准单点注入：只有被分类器判定为高置信度命中的技能，才会在执行当下被动态注入给 Claude Code。 安装仅需一行命令： npx claude-code-templates@latest --mod productivity/jev-skill-suggestion 这种把“工具选择”抽离给轻量分类器、让大模型只专注“任务执行”的架构，正在成为复杂 Agent 系统降本增效的标配设计。随着技能库扩充，未来配合端侧超轻量分类模型（如社区正在探索的 MLX 移植版），整体延迟和隐私表现还会更进一步。 大家目前在 Claude Code 里常驻了多少个自定义技能？这种动态路由机制是否刚好击中了你的上下文焦虑？ #ClaudeCode #AIAgent #AgentSkills #TypeSafeAI #LLM #PromptEngineering #AI编程 #OpenSource #开发者 #独立开发 #全栈开发 #Vercel ❤️ 77 likes on X
- Author
- @realfxw
- Use case
- Context & Memory
- Added
- 2026-09-25
All figures come from the author. Check the source before you quote them.
More in Context & Memory
- morph on X: TypeSafe founder Diogo Almeida introduced Jev, and the field I would inspect fir▲ 0x.com
- Jetwani Avinash on X: First thing I built on Jev: a memory gate for Claude Code.▲ 0x.com0.3s
- Jerry Xu on X: tab-jev: jev-like model + tabular foundation model = an in-context learner for y▲ 0x.com
- GitHub - Cairn-ink/cairn-jev-lab: Test what your AI should remember. An experimental, source-aware memory admission evaluator powered by Jev▲ 0github.com
- GitHub - tamaratran/fast-jev-compaction: Claude Code plugin that replaces the compaction summary with Jev decisions: every tool call and res▲ 0github.com
- GitHub - yangyu666/dsh-jev-prune: Jev-judged context compaction for DeepSeek Harness: semantic tool-result pruning + deterministic receipt c▲ 0github.com
Jev guides for this use case
- How Jev sorts a build into one of 21 use casesThe 21 criteria Jev classifies against, published in full, plus what the reviewer sees and how ambiguity is handled.
- What people actually build with Jev, by use caseAll 21 use cases with live counts from real submissions, plus what each one is actually for.
- Is Jev reliable? The honest limitationsWhat the no-hallucination claim really covers, where the accuracy ceiling sits, and the failure mode we hit ourselves.
Bid history
No bids yet — the first one takes this project straight to the spotlight.