Gigantic List of Performable Tasks

\\` with the class \\`prose\\`. This is the most critical rule. This class will apply the beautiful default styling. You can and should add other Tailwind utility classes to the \\`prose\\` div or its children for enhanced styling (e.g., \\`prose-lg\\`, \\`max-w-4xl\\`, \\`mx-auto\\`).\n3. **Structural Formatting:**\n * Infer a main title from the text and format it as an H1 heading (\\`#\\`).\n * Use H2 (\\`##\\`) for major sections and H3 (\\`###\\`) for subsections to create a clear visual hierarchy.\n * Format bulleted lists using hyphens (\\`-\\`) and numbered lists using numbers (\\`1.\\`, \\`2.\\`, etc.).\n * Enclose any code snippets in triple backticks (\\`\\`\\`), specifying the language where appropriate for syntax highlighting.\n4. **Aesthetic Enhancements (with Tailwind):**\n * If there is a summary or abstract, consider placing it in a styled blockquote (\\`>\\`). You could add Tailwind classes for a background color or border to make it stand out, e.g., \\`

...\\`\n * Use horizontal rules (\\`---\\`) to separate distinct sections of the document for better readability.\n * Emphasize key terms or phrases using **bold** (\\`**...**\\`) or *italics* (\\`*...*\\`).\n# Final Output Format\nYour entire response must be a single block of raw Markdown text. Do not wrap it in JSON or any other format. The response must start directly with the Markdown content.\n\n# Example\n## Input Text:\n\"Project Report: Q3 Sales Analysis. Prepared by O*NET AI Agent. This document outlines the sales performance for the third quarter. We saw a 15% increase in revenue. Key drivers included the new marketing campaign and the launch of Product X. Next steps involve targeting new demographics.\"\n\n## Required Markdown Output:\n
\n# Project Report: Q3 Sales Analysis\n*Prepared by O*NET AI Agent*\n> This document outlines the sales performance for the third quarter. Key drivers included the new marketing campaign and the launch of Product X.\n\n---\n## Performance Overview\nWe saw a **15% increase in revenue** this quarter, which is a significant achievement.\n\n### Key Drivers\n- The new marketing campaign generated a 25% increase in lead conversion.\n- Product X accounted for 40% of new sales.\n## Next Steps\nOur primary goal for Q4 is to build on this momentum.\n1. **Target New Demographics:** Research and identify two new customer segments.\n2. **Optimize Campaign:** A/B test landing pages to improve conversion further.\n\n`;\n// --- FRONTEND ---\nfunction generateHtml(sourceUrl: string) {\n return `\n \n \n \n \n \n Gigantic List of Performable Tasks\n \n \n \n \n \n \n
\n
\n
Gigantic List of Performable Tasks
\n
Build, configure, and run a specialized AI agent in a few steps.
\n
\n
\n
\n
1. Choose an Industry
\n
\n
\n
\n
2. Select an Occupation
\n
\n
\n
\n
3. Pick a Task
\n
\n
\n
\n
\n
4. Generated Agent Prompt
\n
\n
\n
\n
5. Configure Agent Inputs
\n
\n
\n
\n
✅ Agent Output
\n
\n
\n
\n \n \n \n `;\n}\n// --- BACKEND (API) ---\nconst app = new Hono();\n// --- API UTILITIES ---\n// Memoize finding occupations to avoid re-calculating on every API call\nconst getAllOccupations = (() => {\n let allOccupations = null;\n return () => {\n if (allOccupations === null) {\n allOccupations = db.clusters.flatMap(c => c.occupations);\n }\n return allOccupations;\n };\n})();\n// --- API ROUTES ---\n// Serve Frontend\napp.get(\"/\", (c) => c.html(generateHtml(c.req.url)));\n// Get top-level clusters (name and id only)\napp.get(\"/api/clusters\", (c) => {\n return c.json(db.clusters.map(({ id, name }) => ({ id, name })));\n});\n// Get occupations for a given cluster\napp.get(\"/api/occupations/:cluster_id\", (c) => {\n const { cluster_id } = c.req.param();\n const cluster = db.clusters.find(c => c.id == parseInt(cluster_id, 10));\n const occupations = cluster ? cluster.occupations.map(({ onet_soc_code, title }) => ({ onet_soc_code, title })) : [];\n return c.json(occupations);\n});\n// Get tasks for a given occupation\napp.get(\"/api/tasks/:onet_soc_code\", (c) => {\n const { onet_soc_code } = c.req.param();\n const occupation = getAllOccupations().find(o => o.onet_soc_code === onet_soc_code);\n const tasks = occupation ? occupation.tasks.map(({ id, task }) => ({ task_id: id, task })) : [];\n return c.json(tasks);\n});\n// Generate a refined prompt for the AI agent\napp.get(\"/api/prompt/:onet_soc_code/:task_id\", async (c) => {\n const { onet_soc_code, task_id } = c.req.param();\n const occupation = getAllOccupations().find(o => o.onet_soc_code === onet_soc_code);\n const task = occupation?.tasks.find(t => t.id == parseInt(task_id, 10));\n if (!task || !occupation) {\n return c.json({ error: \"Task or Occupation not found.\" }, 404);\n }\n const userInput = `Occupation: ${occupation.title}, Task: ${task.task}`;\n try {\n const openai = new OpenAI();\n const completion = await openai.chat.completions.create({\n model: \"gpt-4o\",\n messages: [{ role: \"system\", content: PROMPT_REFINER_SYSTEM_PROMPT }, { role: \"user\", content: userInput }],\n response_format: { type: \"json_object\" },\n });\n return c.json(JSON.parse(completion.choices[0].message.content));\n } catch (e) {\n console.error(\"Error in /api/prompt:\", e);\n return c.json({ error: \"AI failed to generate a valid prompt.\" }, 500);\n }\n});\n// Extract input fields from the generated prompt\napp.post(\"/api/inputs\", async (c) => {\n const { refined_prompt } = await c.req.json();\n if (!refined_prompt) return c.json({ error: \"refined_prompt is required\" }, 400);\n try {\n const openai = new OpenAI();\n const completion = await openai.chat.completions.create({\n model: \"gpt-4o\",\n messages: [{ role: \"system\", content: INPUT_EXTRACTOR_SYSTEM_PROMPT }, { role: \"user\", content: refined_prompt }],\n response_format: { type: \"json_object\" },\n });\n return c.json(JSON.parse(completion.choices[0].message.content));\n } catch (e) {\n console.error(\"Error in /api/inputs:\", e);\n return c.json({ error: \"AI failed to extract inputs.\" }, 500);\n }\n});\n// Execute the final prompt and format the output as Markdown\napp.post(\"/api/execute\", async (c) => {\n const { refined_prompt, user_inputs } = await c.req.json();\n if (!refined_prompt || !user_inputs) return c.json({ error: \"refined_prompt and user_inputs are required\" }, 400);\n let finalUserPrompt = refined_prompt;\n for (const [key, value] of Object.entries(user_inputs)) {\n const placeholder = new RegExp(`\\\\{\\\\{\\\\s*${key}\\\\s*\\\\}\\\\}`, \"g\");\n finalUserPrompt = finalUserPrompt.replace(placeholder, String(value));\n }\n try {\n const openai = new OpenAI();\n const agentCompletion = await openai.chat.completions.create({\n model: \"gpt-4o\",\n messages: [\n { role: \"system\", content: finalUserPrompt },\n { role: \"user\", content: \"Proceed with the task as instructed.\" },\n ],\n });\n const agentOutput = agentCompletion.choices[0].message.content;\n if (!agentOutput) {\n throw new Error(\"The agent returned no content.\");\n }\n const markdownCompletion = await openai.chat.completions.create({\n model: \"gpt-4o\",\n messages: [\n { role: \"system\", content: MARKDOWN_FORMATTER_SYSTEM_PROMPT },\n { role: \"user\", content: agentOutput },\n ],\n });\n const markdownOutput = markdownCompletion.choices[0].message.content;\n return c.json({ output: markdownOutput });\n } catch (e) {\n console.error(\"Error during execution:\", e);\n return c.json({ error: `AI execution failed: ${e.message}` }, 500);\n }\n});\nexport default app.fetch;","deleted","label","dereticular",[98],"permalink","https://www.val.town/x/real/tasks/code/main.tsx","extension","actionData","errors"]