菜单
Use when implementing or reviewing CADIS model provider traits, provider capabilities, streaming responses, tool-call support, OpenAI, Ollama, Anthropic, Gemini, OpenRouter, LM Studio, or custom HTTP providers.
用 Codex 或 Claude 帮你安装 复制这段 Prompt，粘贴到 Codex、Claude 或其他助手里，让它检查 Skill 页面并帮你完成安装。
基于 SOC 职业分类
Use when updating CADIS contributor docs, release docs, changelog, GitHub issue templates, PR templates, CI, README, governance, license notices, or project planning documents.
Use when implementing or reviewing approval policy, risk classes, sandbox rules, secret redaction, tool permissioning, threat model, or security-sensitive behavior.
Use when changing CADIS protocol messages, event schemas, content routing, HUD protocol, approval payloads, model catalog events, or compatibility rules.
Use when implementing or reviewing CADIS HUD, RamaClaw UI adaptation, config window, agent rename, theme picker, voice/model settings, orbital HUD, approval cards, screenshot parity, or desktop window behavior.
Use when implementing or reviewing CADIS Rust workspace structure, cadisd daemon, cadis CLI, session core, event bus, store, or core crate boundaries.
Use when implementing native CADIS tools such as file, shell, git, patch, browser bridge, tool registry, tool schemas, lifecycle events, timeout, cancellation, or tool result handling.
| name | cadis-model-provider |
| description | Use when implementing or reviewing CADIS model provider traits, provider capabilities, streaming responses, tool-call support, OpenAI, Ollama, Anthropic, Gemini, OpenRouter, LM Studio, or custom HTTP providers. |
docs/01_PRD.mddocs/04_TRD.mddocs/06_IMPLEMENTATION_PLAN.mdtool_calling
vision
json_schema
reasoning_effort
prompt_cache
max_context_tokens
local_model
streaming
Prove one cloud provider and one local provider early. Good defaults are OpenAI and Ollama.
For OpenAI-specific work, also use the openai-docs system skill.