菜单
Reference older Anthropic model-specific reasoning patterns and migrate them to the current Anthropic operations skill. Use when maintaining a legacy Claude 3.7-specific example or translating it to current Claude model selection guidance.
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
Checks for potential drug-drug interactions (DDIs) between a list of medications.
Evaluate and observe agent systems with traced runs, curated datasets, rollout gates, and regression loops. Use when an agent workflow is moving beyond prototype and needs measurable reliability before broader autonomy or production rollout.
Reference older DeepSeek R1-specific reasoning examples and migrate them to current DeepSeek API operations guidance. Use when preserving an open-model reasoning example without treating raw chain-of-thought exposure as a stable product contract.
Reference older Gemini Flash multimodal examples and migrate them to current Gemini operations guidance. Use when a pinned Gemini 2.0 example needs to be translated into the active Google GenAI SDK and model-selection path.
Design a supervisor-worker or debate-style multi-agent coordinator with explicit delegation and synthesis boundaries. Use when a task genuinely benefits from controlled specialization rather than a single-agent loop.
Study OpenAI Swarm as a lightweight historical handoff pattern and migration reference. Use when maintaining an existing Swarm prototype or when you need to translate a small handoff demo into the modern OpenAI Agents SDK.
| name | claude-reasoning-transition-agent |
| description | Reference older Anthropic model-specific reasoning patterns and migrate them to the current Anthropic operations skill. Use when maintaining a legacy Claude 3.7-specific example or translating it to current Claude model selection guidance. |
| measurable_outcome | Convert a model-pinned Claude reasoning example into a provider-current implementation plan with one migration note. |
| allowed-tools | ["read_file","run_shell_command"] |
This directory is retained as a model-specific reference example. It should not be the primary starting point for current Anthropic integration work.
Skills/AI_Providers/Anthropic_Claude_Operations_2026/ for active model and deprecation guidance.