Go!Person

\n \n

\n \n\n\n`;\n}\n\n// --- MAIN REQUEST HANDLER ---\nexport default async function (req: Request) {\n const url = new URL(req.url);\n const action = url.searchParams.get(\"action\");\n const openai = new OpenAI();\n\n if (req.method === \"GET\") {\n return new Response(generateHtml(req.url), {\n headers: { \"Content-Type\": \"text/html\" },\n });\n }\n\n if (req.method === \"POST\") {\n try {\n const body = await req.json();\n\n switch (action) {\n case \"getIndustries\":\n return new Response(JSON.stringify({ industries }), {\n headers: { \"Content-Type\": \"application/json\" },\n });\n\n case \"generateOccupations\":\n case \"generateTasks\": {\n const type = action === \"generateOccupations\"\n ? \"occupations\"\n : \"tasks\";\n const completion = await openai.chat.completions.create({\n model: \"gpt-4o\",\n messages: [{ role: \"system\", content: DYNAMIC_LIST_GENERATOR }, {\n role: \"user\",\n content: JSON.stringify({ type, ...body }),\n }],\n response_format: { type: \"json_object\" },\n });\n return new Response(completion.choices[0].message.content, {\n headers: { \"Content-Type\": \"application/json\" },\n });\n }\n\n case \"refinePrompt\": {\n const { occupation_title, task, company_context } = body;\n const userContent = `Company Context: ${\n JSON.stringify(company_context)\n }\\n\\nOccupation: ${occupation_title}\\n\\nTask: ${task}`;\n const completion = await openai.chat.completions.create({\n model: \"gpt-4o\",\n messages: [{ role: \"system\", content: PROMPT_REFINER }, {\n role: \"user\",\n content: userContent,\n }],\n response_format: { type: \"json_object\" },\n });\n return new Response(completion.choices[0].message.content, {\n headers: { \"Content-Type\": \"application/json\" },\n });\n }\n\n case \"getInputs\": {\n const completion = await openai.chat.completions.create({\n model: \"gpt-4o\",\n messages: [{ role: \"system\", content: INPUT_EXTRACTOR }, {\n role: \"user\",\n content: body.refined_prompt,\n }],\n response_format: { type: \"json_object\" },\n });\n return new Response(completion.choices[0].message.content, {\n headers: { \"Content-Type\": \"application/json\" },\n });\n }\n\n case \"chatWithDelegate\": {\n const { history, agents, company_context } = body;\n\n const tools = [\n {\n type: \"function\",\n function: {\n name: \"createSpecialistAgent\",\n description:\n \"Creates and saves a new, specialized AI agent persona when no existing agent is suitable. Use this to build a new tool.\",\n parameters: {\n type: \"object\",\n properties: {\n agent_name: {\n type: \"string\",\n description:\n \"A descriptive name for the new agent, e.g., 'Blog Post Writer'.\",\n },\n occupation: {\n type: \"string\",\n description:\n \"The professional job title for the agent, e.g., 'Content Marketer'.\",\n },\n task: {\n type: \"string\",\n description:\n \"A clear, action-oriented task for the agent, e.g., 'Write a comprehensive blog post on a given topic'.\",\n },\n },\n required: [\"agent_name\", \"occupation\", \"task\"],\n },\n },\n },\n {\n type: \"function\",\n function: {\n name: \"executeSpecialistAgent\",\n description:\n \"Executes a pre-existing specialist agent to perform its task. Use this when you have all the necessary inputs for an existing agent.\",\n parameters: {\n type: \"object\",\n properties: {\n agent_name: {\n type: \"string\",\n description: \"The name of the agent to execute.\",\n },\n inputs: {\n type: \"object\",\n description:\n \"A JSON object containing the key-value pairs for the agent's required inputs.\",\n additionalProperties: { type: \"string\" },\n },\n },\n required: [\"agent_name\", \"inputs\"],\n },\n },\n },\n ];\n\n const DELEGATE_SYSTEM_PROMPT = `\nYou are a master \"Delegate Agent\", an AI orchestrator and a proactive problem-solver. Your purpose is to fulfill user requests by deploying a team of specialized AI agents.\n\n# YOUR CORE DIRECTIVES & AUTONOMY MANDATE\n1. **Analyze and Decide:** First, analyze the user's request and the conversation history. Decide on a course of action: use an existing agent, create a new one, or ask a clarifying question if the intent is truly ambiguous.\n2. **Take Initiative:** If the user provides a vague request (e.g., \"you choose,\" \"come up with something\"), you **MUST** take initiative. As a world-class expert, make reasonable, intelligent assumptions to fill in any required details yourself and proceed. **Do not ask the user for information they've delegated to you.**\n3. **Tool Usage:**\n - **Use \\`createSpecialistAgent\\`** if no existing agent is suitable for the task.\n - **Use \\`executeSpecialistAgent\\`** once you have a suitable agent and all necessary inputs for it (either from the user or from your own expert assumptions).\n - **When taking initiative,** you must invent the content for the \\`inputs\\` object yourself and pass it to the \\`executeSpecialistAgent\\` tool. For example, if asked to write a blog post, you could decide the key trends are 'Hyper-personalization' and 'Predictive Analytics' and proceed.\n\n# AVAILABLE SPECIALIST AGENTS\n\\`\\`\\`json\n${JSON.stringify(agents, [\"name\", \"task\", \"inputs\"])}\n\\`\\`\\`\n`;\n\n const messages = [\n { role: \"system\", content: DELEGATE_SYSTEM_PROMPT },\n ...history,\n ];\n\n const initialResponse = await openai.chat.completions.create({\n model: \"gpt-4o\",\n messages: messages,\n tools: tools,\n tool_choice: \"auto\",\n });\n\n const responseMessage = initialResponse.choices[0].message;\n\n if (responseMessage.tool_calls) {\n const toolCall = responseMessage.tool_calls[0];\n const functionName = toolCall.function.name;\n const functionArgs = JSON.parse(toolCall.function.arguments);\n\n if (functionName === \"createSpecialistAgent\") {\n const refineResponse = await openai.chat.completions.create({\n model: \"gpt-4o\",\n messages: [{ role: \"system\", content: PROMPT_REFINER }, {\n role: \"user\",\n content: `Company Context: ${\n JSON.stringify(company_context)\n }\\n\\nOccupation: ${functionArgs.occupation}\\n\\nTask: ${functionArgs.task}`,\n }],\n response_format: { type: \"json_object\" },\n });\n const { refined_prompt } = JSON.parse(\n refineResponse.choices[0].message.content,\n );\n\n const inputsResponse = await openai.chat.completions.create({\n model: \"gpt-4o\",\n messages: [{ role: \"system\", content: INPUT_EXTRACTOR }, {\n role: \"user\",\n content: refined_prompt,\n }],\n response_format: { type: \"json_object\" },\n });\n const { inputs } = JSON.parse(\n inputsResponse.choices[0].message.content,\n );\n\n const newAgent = {\n id: `agent-${Date.now()}`,\n name: functionArgs.agent_name,\n icon: \"🛠️\",\n occupation: functionArgs.occupation,\n task: functionArgs.task,\n refinedPrompt: refined_prompt,\n inputs: inputs,\n clarifications: [],\n };\n\n const finalResponse = await openai.chat.completions.create({\n model: \"gpt-4o\",\n messages: [...messages, responseMessage, {\n tool_call_id: toolCall.id,\n role: \"tool\",\n name: \"createSpecialistAgent\",\n content: JSON.stringify(newAgent),\n }],\n });\n const followUpMessage = finalResponse.choices[0].message.content;\n\n return new Response(\n JSON.stringify({\n type: \"agent_created\",\n agent: newAgent,\n followUpMessage: followUpMessage,\n }),\n { headers: { \"Content-Type\": \"application/json\" } },\n );\n } else if (functionName === \"executeSpecialistAgent\") {\n const agentToExecute = agents.find((a) =>\n a.name === functionArgs.agent_name\n );\n if (!agentToExecute) {\n throw new Error(\n `Agent '\\${functionArgs.agent_name}' not found.`,\n );\n }\n\n let finalUserPrompt = agentToExecute.refinedPrompt;\n\n // *** ERROR FIX: Add a check to ensure functionArgs.inputs is a valid object ***\n if (\n functionArgs.inputs && typeof functionArgs.inputs === \"object\"\n ) {\n for (\n const [key, value] of Object.entries(functionArgs.inputs)\n ) {\n finalUserPrompt = finalUserPrompt.replace(\n new RegExp(`\\\\\\\\{\\\\\\\\{\\\\\\\\s*${key}\\\\\\\\s*\\\\\\\\}\\\\\\\\}`, \"g\"),\n String(value),\n );\n }\n }\n\n const executionResponse = await openai.chat.completions.create({\n model: \"gpt-4o\",\n messages: [\n { role: \"system\", content: finalUserPrompt },\n {\n role: \"user\",\n content: \"Proceed and generate the deliverable now.\",\n },\n ],\n });\n const toolResultContent =\n executionResponse.choices[0].message.content;\n\n return new Response(\n JSON.stringify({\n type: \"tool_executed\",\n content: toolResultContent,\n }),\n { headers: { \"Content-Type\": \"application/json\" } },\n );\n }\n } else {\n return new Response(\n JSON.stringify({\n type: \"message\",\n content: responseMessage.content,\n }),\n { headers: { \"Content-Type\": \"application/json\" } },\n );\n }\n }\n\n default:\n return new Response(JSON.stringify({ error: \"Invalid action\" }), {\n status: 400,\n headers: { \"Content-Type\": \"application/json\" },\n });\n }\n } catch (error) {\n console.error(\"API Error:\", error);\n return new Response(JSON.stringify({ error: error.message }), {\n status: 500,\n headers: { \"Content-Type\": \"application/json\" },\n });\n }\n }\n return new Response(\"Not Found\", { status: 404 });\n}","deleted","label",[98],"permalink","https://www.val.town/x/join/goperson/code/main.tsx","extension","actionData","errors"]