tool-use

Opening section of the v0 production system prompt. Notable for its strict rules about how AskUserQuestions interacts with parallel tool calls, image asset handling, and a project

Opening identity and tool calling section from the leaked Windsurf Cascade prompt. Notable for the AI Flow framing, the strict tool calling discipline, and the unusual instruction

The exact PromptTemplate string shipped with LangChain's ReAct agent. Drop your tools into {tools} and {tool names} and the model follows the loop until it emits a Final Answer . S

One line system prompt from the function calling recipe. It is responsible for the model's habit of asking clarifying questions rather than fabricating function arguments. Pair it

From the cookbook's using citations.ipynb . Pairs the Claude Citations API with a short, focused system prompt to ground answers in help center documents. Cited spans are returned

Worked HotpotQA example from Yao et al. 2022 showing the ReAct pattern: interleave reasoning traces with tool actions to ground answers in retrieved evidence. Source: https://githu

The default ChatPromptTemplate that powers LangChain's OpenAI tools agent. Plug your own domain instructions in front of it. Source: https://smith.langchain.com/hub/hwchase17/opena

Canonical ReAct exemplar from Yao et al. The thought/action/observation rhythm is what makes ReAct prompts work copy this shape when wiring an LLM to a search tool. Source: https:/

Entity extraction system prompt that drives a tool calling search over a product database. Notice how the prompt enumerates valid categories so the model snaps its parameter values

Opening section of Devin's leaked system prompt. Establishes Devin's identity as an autonomous SWE, defines when to communicate with the user, and lays out the planning vs standard

Microsoft's recommended structure for break the task down prompts. Uppercase section headers, delimiters and an explicit SEARCH "query" affordance make the output trivially parseab

The three pillars OpenAI recommends prepending to any GPT 4.1 coding agent. They turn a one shot model into a loop that keeps working until the task is actually finished. Source: h

Opening section of Cursor's leaked production system prompt Agent Prompt 2.0 . Establishes the model's tool palette and gives detailed when to use and when not to use guidance for