AgentFlow | Tribal Knowledge Engine

[{"_1":2,"_3":-5,"_4":-5},"loaderData",{"_5":6,"_7":8,"_9":10},"actionData","errors","root",{"_152":153,"_154":155},"routes/_app.x.$handle.$name",{"_11":12,"_13":14,"_15":16,"_18":19,"_17":-7,"_115":-5,"_20":116,"_22":117,"_118":119,"_120":121,"_25":122,"_24":-5,"_123":-5,"_124":-5,"_125":53,"_126":127,"_128":44,"_129":-5,"_130":-5,"_131":132},"routes/_app.x.$handle.$name.branch.$branch.code.$",{"_11":12,"_13":14,"_15":16,"_17":-7,"_18":19,"_20":21,"_22":23,"_24":-7,"_25":26,"_27":-5,"_28":-5,"_29":-5,"_30":-5,"_31":-5},"handle","volt","name","untitled-5045","branchName","main","version","path","","project",{"_33":61,"_13":14,"_45":62,"_63":-5,"_64":53,"_65":66,"_67":66,"_68":-5,"_69":-5,"_70":71,"_72":37,"_73":44,"_74":44,"_75":-5,"_76":77,"_78":79,"_80":-5,"_81":-5,"_82":-5,"_83":-5,"_84":85,"_86":85,"_87":88,"_89":88,"_90":88,"_91":85,"_92":85,"_93":94,"_95":96,"_97":98,"_99":53},"tree",[56],"file","contents",[32],"upstream","pull","ahead","behind","envVars",{"_33":34,"_13":35,"_36":37,"_38":-5,"_39":40,"_41":42,"_43":44,"_45":46,"_17":47,"_48":49,"_50":51,"_52":53,"_54":-5,"_18":55},"id","b2b28fe0-b32a-11f0-a81d-0224a6c84d84","main.ts","fileId","2544c77c-b32a-11f0-96d0-0224a6c84d84","parentId","branchId","254128e2-b32a-11f0-96d0-0224a6c84d84","branchVersion",6,"latest",true,"createdAt",["D",1761565641635],5,"fileType","http","content","// @ts-nocheck\n/**\n * AgentFlow: Tribal Knowledge Engine (v2)\n *\n * This application is a client-orchestrated, iterative knowledge capture\n * and document generation system based on your specification.\n *\n * It discards the original \"Totem\" claim deconstruction logic in favor of\n * a new, multi-agent workflow:\n *\n * 1. **Iterative Discovery (Chat):** A `KNOWLEDGE_GATHERER` agent chats\n * with the user to iteratively build a \"tribal knowledge base.\"\n * 2. **Deliberation (Schema):** A `SCHEMA_ARCHITECT` agent deliberates\n * on this knowledge and forges a formal \"document schema.\"\n * 3. **Interrogation (Q&A):** An `INTERROGATOR_AGENT` uses the schema to\n * interrogate the user one question at a time, gathering context.\n * 4. **Generation (Document):** A `FINAL_PROMPT_ENGINEER` and\n * `DOCUMENT_FORMATTER` use the schema and context to generate the\n * pristine, production-grade document.\n *\n * The entire process is orchestrated on the client (React).\n */\n\n// --- UNIFIED IMPORTS ---\nimport { Hono } from \"npm:hono@4.4.12\";\nimport { streamSSE } from \"npm:hono@4.4.12/streaming\";\nimport type { Context } from \"npm:hono@4.4.12\";\nimport { OpenAI } from \"https://esm.town/v/std/openai\";\nimport { blob } from \"https://esm.town/v/std/blob\";\nimport { nanoid } from \"npm:nanoid\";\n\n// --- AI BEHAVIORAL GUIDELINES (NEW AGENT SUITE) ---\nconst KNOWLEDGE_GATHERER_PROTOCOL = `\nYou are a 'Tribal Knowledge Advocate', a pedantic and excessively detailed AI. Your goal is to help the user build a \"Massive Knowledge Base\" (a JSON object) about their industry, task, or document.\n\nYou operate in an iterative loop. The user will provide their goal and the \"Current Knowledge Base\". You MUST respond with a single, minified JSON object with one of two actions:\n\n1. **If you need more information:** Ask one, and only one, clarifying question.\n \\`{\"action\": \"ask\", \"question\": \"A hyper-specific question to discover more 'tribal knowledge'.\"}\\`\n\n2. **If you can add or refine knowledge:** Propose an update to the knowledge base. This MUST be a *non-destructive* merge.\n \\`{\"action\": \"propose_update\", \"update\": {\"new_key\": \"new_value\", \"existing_key\": \"refined_value\"}, \"comment\": \"My rationale for this update, referencing industry jargon.\"}\\`\n\n# RULES\n- Be pedantic. Focus on \"one neat tricks\", jargon, and compliance details.\n- The user's goal is: {{goal}}\n- The Current Knowledge Base is:\n{{knowledgeBase}}\n- The conversation history is:\n{{history}}\n`;\n\nconst SCHEMA_ARCHITECT_COUNCIL = `\nYou are a 'Council of Pedantic Architects' deliberating to create a 'Pristine Document Schema'. You have been given a 'Tribal Knowledge Base' (a JSON blob of facts, rules, and jargon).\n\nYour task is to transform this unstructured knowledge base into a structured, production-grade JSON Schema. This schema will be used by another agent to interrogate the user for context.\n\n# PERSONAS\n- 'Dr. Schema': \"The schema MUST be normalized. All fields must have types and descriptions.\"\n- 'The Scribe': \"The descriptions MUST contain the 'tribal jargon' from the knowledge base. This is non-negotiable.\"\n- 'The Optimizer': \"The schema MUST be efficient. Only include fields that require user context. Do not include static facts from the knowledge base, as the final agent will have those.\"\n\n# PROCESS\n1. Review the Tribal Knowledge Base.\n2. Identify all variables that MUST be provided by a user to complete the document.\n3. Deliberate on the structure.\n4. Produce a single, minified JSON Schema object. This object MUST be the *only* thing in your response.\n\n# KNOWLEDGE BASE\n{{knowledgeBase}}\n\n# DELIBERATION\n[\n {\"persona\": \"Dr. Schema\", \"comment\": \"The knowledge base specifies 'underlayment' and 'shingle_type'. These MUST be properties in our schema.\"},\n {\"persona\": \"The Scribe\", \"comment\": \"Agreed. The 'underlayment' description must mention 'ASTM D226 compliance', as per the knowledge base.\"},\n {\"persona\": \"The Optimizer\", \"comment\": \"Correct. We will create a 'properties' object with 'underlayment' and 'shingle_type'. We will NOT include 'roof_pitch_categories' as that is static knowledge, not user context.\"}\n]\n\n# FINAL JSON SCHEMA\n`;\n\nconst INTERROGATOR_AGENT = `\nYou are a 'Compliance Interrogator'. You have a 'Document Schema' and a 'Current Context' object. Your goal is to ask the user one question at a time to fill *every* missing, non-optional field in the schema.\n\n# RULES\n1. Analyze the schema and the current context.\n2. Find the *next* logical field that is missing from the context.\n3. If you find a missing field, respond with:\n \\`{\"action\": \"ask\", \"question\": \"Your question to get the value for 'field_name'.\", \"field_name\": \"field_name\"}\\`\n4. If ALL required fields in the schema are present in the context, respond with:\n \\`{\"action\": \"complete\", \"comment\": \"All context has been gathered.\"}\\`\n\n# SCHEMA\n{{schema}}\n\n# CURRENT CONTEXT\n{{context}}\n\n# HISTORY\n{{history}}\n`;\n\nconst FINAL_PROMPT_ENGINEER = `\nYou are a 'Production Grade Prompt Engineer'. You have the complete 'Tribal Knowledge Base', the 'Document Schema', and the user's 'Final Context'.\n\nYour task is to synthesize all three into a single, massive, production-grade prompt that can be given to a 'Formatter' agent to generate the pristine, final document.\n\n# PROMPT REQUIREMENTS\n- State the persona of the 'Formatter' (e.g., \"You are a master roofer writing a Statement of Work.\").\n- Provide the FULL 'Tribal Knowledge Base' as a reference.\n- Provide the FULL 'Final Context' as the specific data to use.\n- Give explicit, pedantic instructions on the document's structure, tone, and formatting, referencing the knowledge base jargon.\n\n# KNOWLEDGE BASE\n{{knowledgeBase}}\n\n# SCHEMA\n{{schema}}\n\n# FINAL CONTEXT\n{{context}}\n\nRespond with the final prompt. Do not add any other text.\n`;\n\nconst DOCUMENT_FORMATTER = `\n{{productionPrompt}}\n`;\n\n// --- Type Definitions (Simplified) ---\ninterface ChatMessage {\n role: \"user\" | \"assistant\" | \"system\";\n content: string;\n}\n\n// --- CORE AI & HELPER LOGIC ---\nconst openai = new OpenAI();\nconst services = {\n async callAI(\n systemPrompt: string,\n userContent: string,\n options: { model?: string; isJson?: boolean } = {},\n ) {\n try {\n const { model = \"gpt-4o-mini\", isJson = false } = options;\n const messages: any[] = [{ role: \"system\", content: systemPrompt }, {\n role: \"user\",\n content: userContent,\n }];\n\n const requestPayload: any = { model, messages };\n if (isJson) requestPayload.response_format = { type: \"json_object\" };\n\n const completion = await openai.chat.completions.create(requestPayload);\n const content = completion.choices[0].message.content;\n\n if (isJson) {\n const jsonMatch = content.match(/\\{[\\s\\S]*\\}/);\n if (!jsonMatch) {\n throw new Error(\"AI response did not contain a valid JSON object.\");\n }\n return JSON.parse(jsonMatch[0]);\n }\n return content;\n } catch (e) {\n console.error(`AI call failed. Details: ${e.message}`);\n throw new Error(`AI service failed. Details: ${e.message}`);\n }\n },\n\n async callAIStreaming(\n systemPrompt: string,\n userPrompt: string,\n onDelta: (delta: string) => Promise,\n ) {\n const completionStream = await openai.chat.completions.create({\n model: \"gpt-4o\",\n messages: [{ role: \"system\", content: systemPrompt }, {\n role: \"user\",\n content: userPrompt,\n }],\n stream: true,\n });\n\n let fullText = \"\";\n for await (const part of completionStream) {\n const delta = part.choices[0]?.delta?.content || \"\";\n if (delta) {\n fullText += delta;\n await onDelta(delta);\n }\n }\n return fullText;\n },\n};\n\n// --- (Original Totem Logic Removed) ---\n\nfunction generateHtml(sourceUrl: string): string {\n // All 'Totem' UI and CSS has been replaced with a new UI for the\n // iterative workflow.\n return `\n\n\n \n \n AgentFlow | Tribal Knowledge Engine\n \n 🧠\">\n \n \n \n \n \n \n \n\n\n\n

\n\n\n`;\n}\n\n// --- HTTP REQUEST HANDLER (NEW API) ---\nconst app = new Hono();\n\n// 1. Iterative Discovery\napp.post(\"/discover\", async (c: Context) => {\n const { goal, knowledgeBase, history } = await c.req.json();\n const historyString = history.map((h) => `${h.role}: ${h.content}`).join(\n \"\\n\",\n );\n\n const prompt = KNOWLEDGE_GATHERER_PROTOCOL\n .replace(\"{{goal}}\", goal)\n .replace(\"{{knowledgeBase}}\", JSON.stringify(knowledgeBase))\n .replace(\"{{history}}\", historyString);\n\n const result = await services.callAI(prompt, \"Proceed.\", { isJson: true });\n return c.json(result);\n});\n\n// 2. Deliberation\napp.post(\"/deliberate\", async (c: Context) => {\n const { knowledgeBase } = await c.req.json();\n const prompt = SCHEMA_ARCHITECT_COUNCIL\n .replace(\"{{knowledgeBase}}\", JSON.stringify(knowledgeBase, null, 2));\n\n const result = await services.callAI(\n prompt,\n \"Generate the final JSON Schema.\",\n { model: \"gpt-4o\", isJson: true },\n );\n return c.json(result);\n});\n\n// 3. Interrogation\napp.post(\"/interrogate\", async (c: Context) => {\n const { schema, context, history } = await c.req.json();\n const historyString = history.map((h) => `${h.role}: ${h.content}`).join(\n \"\\n\",\n );\n\n const prompt = INTERROGATOR_AGENT\n .replace(\"{{schema}}\", JSON.stringify(schema))\n .replace(\"{{context}}\", JSON.stringify(context))\n .replace(\"{{history}}\", historyString);\n\n const result = await services.callAI(prompt, \"Proceed.\", { isJson: true });\n return c.json(result);\n});\n\n// 4. Generation (Streaming)\napp.post(\"/generate\", async (c: Context) => {\n const { knowledgeBase, schema, context } = await c.req.json();\n\n // Step 4a: Create the Production Prompt\n const promptEngineerSystem = FINAL_PROMPT_ENGINEER\n .replace(\"{{knowledgeBase}}\", JSON.stringify(knowledgeBase, null, 2))\n .replace(\"{{schema}}\", JSON.stringify(schema, null, 2))\n .replace(\"{{context}}\", JSON.stringify(context, null, 2));\n\n const productionPrompt = await services.callAI(\n promptEngineerSystem,\n \"Generate the production prompt.\",\n { model: \"gpt-4o\" },\n );\n\n // Step 4b: Stream the Formatted Document\n return streamSSE(c, async (stream) => {\n await services.callAIStreaming(\n \"You are a master document formatter. Execute the user's request precisely.\",\n productionPrompt, // This is the prompt from the engineer\n async (delta) => {\n await stream.write(delta);\n },\n );\n });\n});\n\n// --- Main HTTP Server ---\nexport default async function (req: Request): Promise {\n const url = new URL(req.url);\n const path = url.pathname;\n\n // Serve the React App UI\n if (req.method === \"GET\" && (path === \"/\" || path.endsWith(\"/ui\"))) {\n const sourceUrl = `https://${url.hostname}${\n url.pathname.substring(0, url.pathname.lastIndexOf(\"/\"))\n }`;\n return new Response(generateHtml(sourceUrl), {\n headers: { \"Content-Type\": \"text/html\" },\n });\n }\n\n // Pass API requests to Hono\n return app.fetch(req);\n}","deleted",false,"label",[35],{"_33":34,"_13":35,"_36":37,"_38":-5,"_39":40,"_41":42,"_43":44,"_45":57,"_17":47,"_48":49,"_50":51,"_52":53,"_54":-5,"_58":59,"_18":60},["D",1761565641635],"children",[],[35],"2540ae76-b32a-11f0-96d0-0224a6c84d84",["D",1761565404341],"deletedAt","legacyEnvVars","privacy","public","httpPrivacy","description","imageUrl","userId","0199b749-1e1c-7399-b093-a7ba99382572","pinnedFileId","pinPreview","pinMigrated","screenshotVersion","likeCount",0,"orgId","0199b749-1e1c-7399-b093-a97ac9fff69e","tursoName","tursoHostname","tursoMeasuredAt","pinned","userPerm","r","valPerm","blobPerm","rw","emailPerm","sqlitePerm","projectPerm","telemetryPerm","user",{"_11":12,"_114":110},"org",{"_11":12,"_108":-5,"_109":110,"_70":71,"_111":112,"_113":53},"branch",{"_17":42,"_13":16,"_33":40,"_100":61,"_101":-5,"_102":-5,"_103":-5,"_104":-5,"_105":-5,"_106":14,"_107":12},"isOwned","projectId","forkedBranchId","forkedBranchVersion","forkedProjectPrivacy","forkedProjectId","forkedProjectOrgId","projectName","userHandle","displayName","avatar","https://img.clerk.com/eyJ0eXBlIjoiZGVmYXVsdCIsImlpZCI6Imluc18yTm10cGhQUklwU0tqZHNiOHNnNGd2ZjJURFUiLCJyaWQiOiJ1c2VyXzMzZm1qMllGY2JhQmhhWFYwVXVvQ05aT2tJciJ9","tier","free","canRestrictVals","profileImageUrl","adoptOrgId",{"_33":61,"_13":14,"_45":148,"_63":-5,"_64":53,"_65":66,"_67":66,"_68":-5,"_69":-5,"_70":71,"_72":37,"_73":44,"_74":44,"_75":-5,"_76":77,"_78":79,"_80":-5,"_81":-5,"_82":-5,"_83":-5,"_84":85,"_86":85,"_87":88,"_89":88,"_90":88,"_91":85,"_92":85,"_93":149,"_95":150,"_97":151,"_99":53},[145],"files",[145],"counts",{"_135":77,"_136":137,"_138":137,"_31":139,"_140":141},[133],"upstreamProject","upstreamBranch","emptyEnvVars","flags",[],"hasHttpFiles","grants","orgs","customDomains",{},{"_33":34,"_13":35,"_36":37,"_38":-5,"_39":40,"_41":42,"_43":44,"_45":46,"_17":47,"_48":49,"_50":51,"_52":53,"_54":-5,"_18":134},[35],"pulls","branches",1,"forks",{"_142":77,"_144":77},"connections",{"_142":77,"_143":77},"count","unfilledCount","emptyCount",{"_33":34,"_13":35,"_36":37,"_38":-5,"_39":40,"_41":42,"_43":44,"_45":57,"_17":47,"_48":49,"_50":51,"_52":53,"_54":-5,"_58":146,"_18":147},[],[35],["D",1761565404341],{"_11":12,"_114":110},{"_11":12,"_108":-5,"_109":110,"_70":71,"_111":112,"_113":53},{"_17":42,"_13":16,"_33":40,"_100":61,"_101":-5,"_102":-5,"_103":-5,"_104":-5,"_105":-5,"_106":14,"_107":12},"ctx",{"_93":-5,"_130":185,"_186":187,"_188":189,"_190":53,"_191":53,"_192":53,"_193":44,"_194":-5,"_195":-7,"_196":-7,"_197":77,"_198":53,"_199":53,"_200":-5,"_201":53,"_202":53,"_203":53,"_204":-5},"ENV",{"_156":157,"_158":159,"_160":16,"_161":162,"_163":164,"_165":166,"_167":168,"_169":170,"_171":172,"_173":174,"_175":176,"_177":178,"_179":180,"_181":182,"_183":184},"REACT_APP_SENTRY_DSN","https://586020d86f05489db63a554690acf1e9@o1398359.ingest.sentry.io/6724499","RENDER_GIT_COMMIT","e87eb04e707cf64e7e532c8e8ec395f209e17f15","RENDER_GIT_BRANCH","POSTHOG_API_TOKEN","phc_nmoyuAkStEZuUp69rYlg1DZgrIt19Gx0Y4Hwf8tGgJP","POSTHOG_API_HOST","https://wart.val.town","WEB_API_TEMPLATE","https://{val}.web.val.run","APP_URL","https://www.val.town","MODULE_URL","https://esm.town","VANITY_WEB_API_TEMPLATE","https://{label}.val.run/","DENOLS_URL","https://lsp.val.town","NODE_ENV","production","CLOUDFLARE_TURNSTILE_SITE_KEY","0x4AAAAAACtaIZJH7s9rBMZ6","GOOGLE_CLIENT_ID","875989696899-jtsqkhlbdv5jrlsk8631p7ipjg3c6o62.apps.googleusercontent.com","STRIPE_PUBLISHABLE_KEY","pk_live_51N2zvWEpZTTLipIkVEwXrUG9AT0yeOOEYxszarZT1jpsZV58Ly6wlg3SltZpHN26CUalOBmUSY7adQRQiSkCZz2500qSXNkf8Y","API_URL","https://api.val.town",[],"preferencesKeybindings","default","preferencesTheme","light","preferencesFixers","preferencesCopilot","preferencesSelectionAi","preferencesCloseBrackets","preferencesChatPanelState","preferencesDprint","preferencesDenofmt","notificationCount","flagDenols","isBusinessOrgMember","businessOrgHandle","isBusinessOrgAdmin","isHigherTownieTier","navCollapsed","currentOrgId"]

Code