omni-settings

Agent Skills › cheaper-inference/OmniRoute › omni-settings

omni-settings

GitHub

提供全局应用设置的读写能力，涵盖系统提示、IP过滤、负载规则及认证配置。支持通过API管理记忆、向量检索及压缩等核心参数，适用于应用初始化与运行时动态调整。

skills/omni-settings/SKILL.md cheaper-inference/OmniRoute

触发场景

需要读取或修改系统级配置配置IP白名单或黑名单调整LLM上下文或压缩策略设置全局默认行为

安装

npx skills add cheaper-inference/OmniRoute --skill omni-settings -g -y

更多选项

不安装直接使用

npx skills use cheaper-inference/OmniRoute@omni-settings

指定 Agent (Claude Code)

npx skills add cheaper-inference/OmniRoute --skill omni-settings -a claude-code -g -y

安装 repo 全部 skill

npx skills add cheaper-inference/OmniRoute --all -g -y

预览 repo 内 skill

npx skills add cheaper-inference/OmniRoute --list

SKILL.md

Frontmatter

{
    "name": "omni-settings",
    "description": "Read and update global application settings: system prompts, thinking budget, IP filters, payload rules, combo defaults, and require-login configuration."
}

Overview

Read and update global application settings: system prompts, thinking budget, IP filters, payload rules, combo defaults, and require-login configuration.

Authentication

All requests require a valid Bearer token or session cookie. Obtain a token via POST /api/auth/login or configure REQUIRE_API_KEY=false for local development.

Endpoints

Payloads

See the full OpenAPI specification at GET /api/openapi/spec or docs/openapi.yaml for detailed request/response schemas.

GET /api/settings/thinking-budget — behavior

Returns proxy-level thinking/reasoning request rewrite settings:

Field	Meaning
`mode`	`passthrough` (leave client reasoning alone — required for Codex visible thinking), `auto` (strips all client thinking fields), `custom`, `adaptive`
`customBudget`	Fixed budget when `mode=custom`
`effortLevel`	Base effort when `mode=adaptive`

Not compression and not "decrypt encrypted reasoning". Full guide: docs/guides/THINKING_BUDGET.md.

Example — keep client-controlled reasoning (Codex/Desktop):

curl -X PUT https://localhost:20128/api/settings/thinking-budget \
  -H "Authorization: Bearer $OMNIROUTE_TOKEN" \
  -H "Content-Type: application/json" \
  -d '{"mode":"passthrough","customBudget":10240,"effortLevel":"medium"}'

Warning: mode=auto deletes reasoning / reasoning_effort / Claude thinking from the outbound body before upstream. That can empty thinking panels even when the client requested Ultra + summary.

版本历史

5dbd2eb 当前 2026-11-03 07:32
d804179 2026-10-19 14:22
新增按需运行的存储完整性检查接口；优化性能，移除每次读取设置时的完整性检查开销；新增选择性提供商账户转移相关API。
e2dc851 2026-09-22 19:11
ba597b6 2026-09-09 13:21
无内容变更
0ad0331 2026-08-28 16:40
3d7ed7a 2026-08-20 06:20
修复因手动编辑SKILL.md导致的生成器同步失败问题，将自定义内容移至保留块以确保持久化。
1cafd32 2026-07-25 11:48

元信息

文件数: 0
版本: 5dbd2eb
Hash: c8d7e548
收录时间: 2026-07-25 11:48

omni-settings

omni-settings

触发场景

安装

SKILL.md

Overview

Authentication

Endpoints

Payloads

GET /api/settings/thinking-budget — behavior

版本历史

同 Skill 集合

元信息

文库