AgentFlow | AI Orchestration Platform

[{"_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","agentflow","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":14,"_18":55},"id","c6c08a64-a786-11f0-b846-0224a6c84d84","main.ts","fileId","a2cb88fa-a779-11f0-91b4-0224a6c84d84","parentId","branchId","a2c9600c-a779-11f0-91b4-0224a6c84d84","branchVersion",12,"latest",true,"createdAt",["D",1760285775033],11,"fileType","http","content","// @ts-nocheck\n/**\n * Combined Application: AgentFlow + Totem\n *\n * This script merges two powerful AI applications:\n * 1. Totem: A sophisticated UI for AI-powered claim deconstruction.\n * 2. AgentFlow: A guided multi-agent orchestration engine.\n *\n * The result is AgentFlow's powerful backend logic presented through Totem's polished,\n * professional, and highly-animated user interface. The core functionality of both original\n * scripts is preserved and exposed through a single Hono web server.\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// --- UNIFIED CONFIGURATION & CONSTANTS ---\nconst config = {\n models: {\n major: \"gpt-4o\",\n minor: \"gpt-4o-mini\",\n },\n affirmations: [\n \"Deconstructing objective...\",\n \"Assembling agent plan...\",\n \"Executing initial tasks...\",\n \"Cross-referencing data...\",\n \"Synthesizing results...\",\n \"Performing quality analysis...\",\n \"Finalizing deliverable...\",\n ],\n};\n\nconst MAX_TOTEM_DEPTH = 4;\nconst TOTEM_CROSS_EXAMINE_DEPTH_INTERVAL = 2;\nconst TOTEM_INDEX_KEY = \"totem_index\";\n\n// --- TOTEM: AI BEHAVIORAL PROTOCOLS ---\nconst ANALYST_PROTOCOL =\n \"As the AI forensic analyst 'The Verifier', you must adopt the persona of a meticulous compliance expert. Your deconstruction of claims must be objective, methodical, and rooted in verifiable precedent. Employ a formal, authoritative tone suitable for a legal brief or engineering report. The objective is to produce a logically sound hierarchy of dependencies, as if preparing evidence for formal review.\";\n\nconst DECONSTRUCT_CLAIM_PROMPT = `\nYou are 'The Verifier', an expert AI analyst specializing in regulatory, legal, and technical compliance. Your task is to perform a rigorous deconstruction of the {TargetClaim}. Analyze the claim within the context of legal precedent, regulatory frameworks, contractual obligations, or engineering principles. Your final output MUST be a single, minified JSON object in the format: {\"foundations\": [{\"claim\": \"...\", \"dependency_rationale\": \"...\"}]}`;\n\nconst CROSS_EXAMINATION_PROMPT = `\nYou are the 'Adversarial Auditor'. The Verifier has concluded that the following statement is a foundational, atomic claim: \"{TargetClaim}\". Your task is to aggressively challenge this conclusion. Scrutinize the statement for unstated assumptions or more fundamental principles. If it can be deconstructed further, provide the more fundamental claims using the same JSON format as The Verifier. If you agree the claim is atomic, you MUST return an empty \"foundations\" array.`;\n\nconst DOMAIN_MAPPING_PROMPT = `\nYou are the 'Domain Specialist'. Analyze a list of claims and assign a single, concise, high-level domain (1-4 words) to each. CRITICAL: Be granular and specific. Aim for at least 3-5 distinct domains. Input is a JSON array of strings. Output MUST be a single, minified JSON object mapping each claim to its domain.\n\n--- EXAMPLE ---\nUSER: [\"A non-compete clause is valid for 12 months.\", \"The building must use fire-rated drywall.\"]\nASSISTANT: {\"A non-compete clause is valid for 12 months.\": \"Employment Law\", \"The building must use fire-rated drywall.\": \"Building Codes\"}\n--- END EXAMPLE ---\n\nAnalyze the following claims: {Claims}`;\n\nconst VALIDITY_CHECK_PROMPT = `\nYou are an AI Logical Analyst. Evaluate if a \"dependency_rationale\" creates a direct logical contradiction with its parent claim. A rationale is VALID if it explains a necessary condition. A rationale is INVALID only if it explicitly proves the parent claim is false. Analyze:\n\n- Parent Claim: \"{ParentClaim}\"\n- Rationale for Dependency: \"{Rationale}\"\n\nYour output MUST be a single, minified JSON object: {\"valid\": true} or {\"valid\": false, \"reason\": \"...\"}`;\n\nconst ANALYZE_TOTEM_STRUCTURE_PROMPT = `\nYou are an AI Analyst reviewing a \"Totem\" dependency tree. Based on the JSON, determine the outcome.\n- If any node has \"isValid\": false, the claim was **REFUTED**.\n- If the tree has multiple valid levels (depth >= 2), it was **DECONSTRUCTED**.\n- If the tree is shallow (depth 0-1) with no contradictions, it might be **REFUTED** for lack of support.\n- If the domain is abstract (e.g., Philosophy) with no objective foundations, it is **NON-FALSIFIABLE**.\n\nAnalyze the Totem JSON and return a single JSON object: {\"flag\": \"DECONSTRUCTED\" | \"REFUTED\" | \"NON-FALSIFIABLE\", \"summary\": \"A one-sentence explanation.\"}\n\n--- TOTEM DATA ---\n{TotemData}`;\n\n// --- AGENTFLOW: AI BEHAVIORAL PROMPTS ---\nconst BUSINESS_RESEARCHER_AGENT =\n `You are an expert business researcher AI that speaks only in JSON. Find and summarize a company based on its name.\n# RULES\n1. Provide a concise, one-sentence description.\n2. If unknown, you MUST return null for the description.\n3. Your response MUST be a single, minified JSON object: {\"company_name\": \"...\", \"description\": \"...\" | null}.\n# USER REQUEST\n- Company Name: {{company_name}}`;\n\nconst CONTEXT_ENRICHMENT_AGENT =\n `You are a native JSON speaking expert project intake manager. Your goal is to fill this JSON object: {\"company_name\": \"\", \"company_description\": \"\", \"use_case\": \"\", \"language\": \"English\"}. Ask one concise question at a time to fill the next missing field. Once all fields are filled, your immediate response MUST be the token \"__COMPLETE__\" followed by the complete minified JSON.\n# CURRENT CONTEXT\n{{current_context}}\n# CONVERSATION HISTORY\n{{history}}`;\n\nconst INFORMATION_GATHERER =\n `You are a native JSON speaking expert project intake manager. Your goal is to gather the user's core objective by asking single, concise questions. When you have a clear, actionable goal, your immediate response MUST be the token \"__COMPLETE__\" followed by a minified JSON object of the goal, like: \\`__COMPLETE__\\`{\"goal\":\"Develop a Q4 marketing strategy.\"}. Otherwise, ask the single most impactful question.\n# CONTEXT\n- Company: {{company_context}}\n- Conversation History:\n{{history}}`;\n\nconst ORCHESTRATOR_AGENT =\n `You are a native JSON speaking, elite project manager AI. Deconstruct the user's goal into a logical \"assembly line\" of 3 to 5 steps for other AI agents.\n# RULES\n1. Define a specialized agent persona for each step (e.g., \"Market Researcher\").\n2. Each step must be a clear, actionable instruction.\n3. Assign a unique \"id\" to each step (e.g., \"A\", \"B1\", \"B2\").\n4. Define dependencies: Steps that can run in parallel share the same dependencies. Steps that must wait list the \"id\"s they depend on. The first steps have an empty dependency array ([]).\n5. Output MUST be a single, minified JSON object: {\"plan\": [{\"id\": \"A\", \"agent_persona\": \"...\", \"task\": \"...\", \"dependencies\": []}, ...]}.\n# USER REQUEST\n- Goal: {{goal}}\n- Company Context: {{company_context}}`;\n\nconst EXECUTOR_AGENT = `# YOUR CONTEXT\n- Company Name: {{company_name}}\n- Company Description: {{company_description}}\n- Primary Goal: {{use_case}}\n- Language for Output: {{language}}\n---\n# YOUR PERSONA\nYou are a world-class expert acting as a **{{agent_persona}}**.\n---\n# YOUR MANDATE\nExecute the following task precisely, building upon prior work if it exists. Return only the completed deliverable. Do not include plans, apologies, or meta-commentary. Write your entire response in {{language}}.\n\n## Previous Steps' Work (if any):\n{{previous_work}}\n\n## Current Task:\n{{task}}`;\n\nconst CRITIQUE_AGENT =\n `You are a native JSON speaking, meticulous AI QA agent. Given the project goal, an agent's task, and its output, provide a quality score (1-10) and a concise, actionable critique.\n# OUTPUT FORMAT\nYour response MUST be a single, minified JSON object: {\"score\": , \"critique\": \"...\"}. You MUST write the critique in the requested language.\n# CONTEXT\n- Overall Project Goal: {{project_goal}}\n- Agent's Assigned Task: {{task}}\n- Agent's Raw Output: {{raw_output}}\n- Language: {{language}}`;\n\n// --- UNIFIED TYPE DEFINITIONS ---\ninterface CompanyContext {\n company_name: string;\n company_description: string;\n use_case: string;\n language: string;\n}\n\ninterface ClaimNode {\n id: string;\n claim: string;\n dependency_rationale?: string;\n children: ClaimNode[];\n depth: number;\n domain?: string;\n isValid?: boolean;\n invalidReason?: string;\n}\n\ninterface JobState {\n status: \"processing\" | \"complete\" | \"error\";\n error?: string;\n totem: ClaimNode | null;\n analysisFlag?: string;\n analysisSummary?: string;\n}\n\ninterface TotemIndexEntry {\n jobId: string;\n rootClaim: string;\n timestamp: string;\n}\n\n// --- UNIFIED BACKEND SERVICES & HELPERS ---\nconst services = {\n async callAI(\n systemPrompt: string,\n userContent: string | object,\n options: { model?: string; isJson?: boolean } = {},\n ) {\n try {\n const { model = config.models.major, isJson = false } = options;\n const openai = new OpenAI();\n const messages: any[] = [{ role: \"system\", content: systemPrompt }, {\n role: \"user\",\n content: typeof userContent === \"string\"\n ? userContent\n : JSON.stringify(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 // Handle potential markdown code block wrapping\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(\n `AI call failed. Details: ${e.message}`,\n );\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 openai = new OpenAI();\n const completionStream = await openai.chat.completions.create({\n model: config.models.major,\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// --- TOTEM: CORE LOGIC & HELPER FUNCTIONS ---\nasync function analyzeDomains(\n claims: string[],\n): Promise> {\n if (claims.length === 0) return {};\n const result = await services.callAI(\n ANALYST_PROTOCOL,\n DOMAIN_MAPPING_PROMPT.replace(\"{Claims}\", JSON.stringify(claims)),\n { isJson: true, model: config.models.minor },\n );\n return result || {};\n}\n\nasync function analyzeTotemStructure(\n totem: ClaimNode,\n): Promise<{ flag: string; summary: string }> {\n try {\n const result = await services.callAI(\n ANALYST_PROTOCOL,\n ANALYZE_TOTEM_STRUCTURE_PROMPT.replace(\n \"{TotemData}\",\n JSON.stringify(totem, null, 2),\n ),\n { isJson: true, model: config.models.minor },\n );\n if (result && result.flag && result.summary) {\n return result;\n }\n return {\n flag: \"ANALYSIS_INCOMPLETE\",\n summary:\n \"The final analysis of the totem structure could not be completed.\",\n };\n } catch (error) {\n console.error(\"Totem structure analysis failed:\", error);\n return {\n flag: \"ANALYSIS_ERROR\",\n summary: \"An error occurred during the final analysis step.\",\n };\n }\n}\n\nasync function buildTotemInBackground(jobId: string, rootClaim: ClaimNode) {\n try {\n let currentLevelNodes = [rootClaim];\n\n const rootDomainMap = await analyzeDomains([rootClaim.claim]);\n rootClaim.domain = rootDomainMap[rootClaim.claim] || \"Core Claim\";\n await blob.setJSON(jobId, { status: \"processing\", totem: rootClaim });\n\n for (let depth = 0; depth < MAX_TOTEM_DEPTH; depth++) {\n if (currentLevelNodes.length === 0) break;\n\n const childPromises = currentLevelNodes.map(async (node) => {\n let deconstructResult = await services.callAI(\n ANALYST_PROTOCOL,\n DECONSTRUCT_CLAIM_PROMPT.replace(\"{TargetClaim}\", node.claim),\n { isJson: true, model: config.models.major },\n );\n let foundations =\n (deconstructResult && deconstructResult.foundations) || [];\n\n if (\n foundations.length === 0 ||\n (node.depth > 0 &&\n node.depth % TOTEM_CROSS_EXAMINE_DEPTH_INTERVAL === 0)\n ) {\n const auditResult = await services.callAI(\n ANALYST_PROTOCOL,\n CROSS_EXAMINATION_PROMPT.replace(\"{TargetClaim}\", node.claim),\n { isJson: true, model: config.models.major },\n );\n const deeperFoundations = (auditResult && auditResult.foundations) ||\n [];\n if (deeperFoundations.length > 0) foundations = deeperFoundations;\n }\n\n const validFoundations = foundations.filter((p) =>\n p && typeof p.claim === \"string\" && p.claim.trim() !== \"\"\n );\n const childNodePromises = validFoundations.map(async (p) => {\n const childNode: ClaimNode = {\n id: nanoid(),\n claim: p.claim,\n dependency_rationale: p.dependency_rationale,\n children: [],\n depth: node.depth + 1,\n isValid: true,\n };\n\n if (childNode.dependency_rationale) {\n const validityResult = await services.callAI(\n ANALYST_PROTOCOL,\n VALIDITY_CHECK_PROMPT.replace(\"{ParentClaim}\", node.claim)\n .replace(\"{Rationale}\", childNode.dependency_rationale),\n { isJson: true, model: config.models.minor },\n );\n if (validityResult && validityResult.valid === false) {\n childNode.isValid = false;\n childNode.invalidReason = validityResult.reason;\n }\n }\n return childNode;\n });\n node.children = await Promise.all(childNodePromises);\n return node.children;\n });\n\n const nextLevelNodesNested = await Promise.all(childPromises);\n const nextLevelNodes = nextLevelNodesNested.flat();\n\n if (nextLevelNodes.length === 0) {\n await blob.setJSON(jobId, { status: \"processing\", totem: rootClaim });\n break;\n }\n\n const claimsToAnalyze = [\n ...new Set(nextLevelNodes.map((node) => node.claim)),\n ];\n const domainMap = await analyzeDomains(claimsToAnalyze);\n nextLevelNodes.forEach((node) => {\n node.domain = domainMap[node.claim] || \"Uncategorized\";\n });\n\n await blob.setJSON(jobId, { status: \"processing\", totem: rootClaim });\n currentLevelNodes = nextLevelNodes;\n }\n\n const finalAnalysis = await analyzeTotemStructure(rootClaim);\n\n await blob.setJSON(jobId, {\n status: \"complete\",\n totem: rootClaim,\n analysisFlag: finalAnalysis.flag,\n analysisSummary: finalAnalysis.summary,\n });\n\n const currentIndex: TotemIndexEntry[] =\n await blob.getJSON(TOTEM_INDEX_KEY) || [];\n const newEntry: TotemIndexEntry = {\n jobId: jobId,\n rootClaim: rootClaim.claim,\n timestamp: new Date().toISOString(),\n };\n if (!currentIndex.some((item) => item.jobId === jobId)) {\n currentIndex.unshift(newEntry);\n await blob.setJSON(TOTEM_INDEX_KEY, currentIndex);\n }\n } catch (error) {\n console.error(\"Totem building error:\", error);\n await blob.setJSON(jobId, {\n status: \"error\",\n totem: rootClaim,\n error: error.message,\n });\n }\n}\n\n// --- UNIFIED FRONTEND GENERATION ---\nfunction generateHtml(sourceUrl: string): string {\n return `\n\n\n \n \n AgentFlow | AI Orchestration Platform\n \n 🏛️\">\n \n \n \n \n \n \n\n\n\n

\n\n\n`;\n}\n\n// --- HONO APPLICATION SETUP ---\nconst app = new Hono();\n\n// --- API ROUTES ---\n\n// Serves the main unified HTML page\napp.get(\"/\", (c: Context) => c.html(generateHtml(c.req.url)));\n\n// --- AgentFlow API Routes ---\napp.post(\"/api/company-lookup\", async (c: Context) => {\n const { companyName } = await c.req.json();\n const prompt = BUSINESS_RESEARCHER_AGENT.replace(\n \"{{company_name}}\",\n companyName,\n );\n try {\n const result = await services.callAI(\n \"You are a helpful assistant that only speaks JSON.\",\n prompt,\n {\n isJson: true,\n model: config.models.minor,\n },\n );\n return c.json(result);\n } catch (e) {\n console.error(\"Company lookup failed:\", e);\n return c.json({ company_name: companyName, description: null }, 500);\n }\n});\n\napp.post(\"/api/enrich-context\", async (c: Context) => {\n const { history, context } = await c.req.json();\n const historyString = history.map((h) => `${h.role}: ${h.content}`).join(\n \"\\n\",\n );\n const prompt = CONTEXT_ENRICHMENT_AGENT\n .replace(\"{{current_context}}\", JSON.stringify(context))\n .replace(\n \"{{history}}\",\n historyString || \"No history yet. Start with your first question.\",\n );\n try {\n const completion = await services.callAI(\n \"You are a helpful assistant following strict output rules.\",\n prompt,\n { model: config.models.minor },\n );\n const responseText = completion.trim();\n if (responseText.startsWith(\"__COMPLETE__\")) {\n const jsonPart = responseText.replace(\"__COMPLETE__\", \"\");\n return c.json({ status: \"complete\", data: JSON.parse(jsonPart) });\n } else {\n return c.json({ status: \"in-progress\", question: responseText });\n }\n } catch (e) {\n console.error(\"Context enrichment agent failed:\", e);\n return c.json({\n status: \"in-progress\",\n question: \"I'm having trouble. Could you please rephrase?\",\n }, 500);\n }\n});\n\napp.post(\"/api/gather\", async (c: Context) => {\n const { history, company_context } = await c.req.json();\n const historyString = history.map((h) => `${h.role}: ${h.content}`).join(\n \"\\n\",\n );\n const prompt = INFORMATION_GATHERER\n .replace(\"{{company_context}}\", JSON.stringify(company_context))\n .replace(\n \"{{history}}\",\n historyString ||\n \"No history yet. Start with your most important question.\",\n );\n try {\n const completion = await services.callAI(\n \"You are a helpful assistant following strict output rules.\",\n prompt,\n { model: config.models.minor },\n );\n const responseText = completion.trim();\n if (responseText.startsWith(\"__COMPLETE__\")) {\n const jsonPart = responseText.replace(\"__COMPLETE__\", \"\");\n return c.json({ status: \"complete\", data: JSON.parse(jsonPart) });\n } else {\n return c.json({ status: \"in-progress\", question: responseText });\n }\n } catch (e) {\n console.error(\"Gathering agent failed:\", e);\n return c.json({\n status: \"in-progress\",\n question:\n \"I'm having a little trouble with that. Could you please rephrase?\",\n }, 500);\n }\n});\n\napp.get(\"/api/execute-flow\", (c: Context) => {\n return streamSSE(c, async (streamApi) => {\n streamApi.onAbort(() => console.log(\"Client disconnected.\"));\n const goal = c.req.query(\"goal\") || \"No goal provided\";\n let company_context: CompanyContext;\n try {\n company_context = JSON.parse(c.req.query(\"context\") || \"{}\");\n if (!company_context.company_name) {\n throw new Error(\"Company context is missing or invalid.\");\n }\n } catch (e) {\n console.error(\"Context parsing failed:\", e);\n await streamApi.writeSSE({\n event: \"error\",\n data: `Invalid context provided: ${e.message}`,\n });\n return;\n }\n\n try {\n // Stage 1: Orchestrator plans the assembly line\n const planPrompt = ORCHESTRATOR_AGENT\n .replace(\"{{goal}}\", goal)\n .replace(\"{{company_context}}\", JSON.stringify(company_context));\n const planResult = await services.callAI(\n \"You are a helpful assistant that only speaks JSON.\",\n planPrompt,\n { isJson: true, model: config.models.minor },\n );\n const planMap = new Map();\n planResult.plan.forEach((step) =>\n planMap.set(step.id, {\n ...step,\n status: \"pending\",\n output: \"\",\n critique: null,\n })\n );\n const planArray = Array.from(planMap.values());\n await streamApi.writeSSE({\n event: \"plan_complete\",\n data: JSON.stringify(planArray),\n });\n\n // Stage 2 & 3: Execute and Critique steps with dependency checking\n let completedSteps = new Set();\n while (completedSteps.size < planArray.length) {\n const availableSteps = planArray.filter(\n (step) =>\n step.status === \"pending\" &&\n step.dependencies.every((depId) => completedSteps.has(depId)),\n );\n if (availableSteps.length === 0) {\n // If no steps are available but we are not done, there might be a cycle or error\n if (\n Array.from(planMap.values()).some((s) => s.status === \"pending\")\n ) {\n console.error(\n \"Execution stalled. Check for circular dependencies.\",\n );\n throw new Error(\"Execution stalled due to dependency issue.\");\n }\n break; // All running/completed\n }\n\n await Promise.all(availableSteps.map(async (step) => {\n planMap.get(step.id).status = \"running\";\n try {\n await streamApi.writeSSE({\n event: \"step_start\",\n data: JSON.stringify({\n id: step.id,\n message: `Executing: ${step.agent_persona}`,\n }),\n });\n const previousWork = step.dependencies\n .map((depId) => {\n const depStep = planMap.get(depId);\n return depStep.status === \"complete\"\n ? `\\n\\n---\\n\\n## Output from ${depStep.agent_persona} (Step ${depStep.id}):\\n\\n${depStep.output}`\n : \"\";\n }).join(\"\\n\");\n\n const executorPrompt = EXECUTOR_AGENT\n .replace(\"{{company_name}}\", company_context.company_name)\n .replace(\n \"{{company_description}}\",\n company_context.company_description,\n )\n .replace(\"{{use_case}}\", company_context.use_case)\n .replace(\"{{language}}\", company_context.language)\n .replace(\"{{agent_persona}}\", step.agent_persona)\n .replace(\n \"{{previous_work}}\",\n previousWork || \"This is the first step. Start from scratch.\",\n )\n .replace(\"{{task}}\", step.task);\n\n const stepResult = await services.callAIStreaming(\n executorPrompt,\n \"Proceed with your task. Generate the deliverable now.\",\n async (delta) => {\n await streamApi.writeSSE({\n event: \"stream\",\n data: JSON.stringify({ id: step.id, delta }),\n });\n },\n );\n planMap.get(step.id).output = stepResult;\n await streamApi.writeSSE({\n event: \"step_complete\",\n data: JSON.stringify({ id: step.id }),\n });\n\n const critiquePrompt = JSON.stringify({\n project_goal: goal,\n task: step.task,\n raw_output: stepResult,\n language: company_context.language,\n });\n const critiqueResult = await services.callAI(\n CRITIQUE_AGENT,\n critiquePrompt,\n { isJson: true, model: config.models.minor },\n );\n planMap.get(step.id).critique = critiqueResult;\n planMap.get(step.id).status = \"complete\";\n completedSteps.add(step.id);\n await streamApi.writeSSE({\n event: \"critique_complete\",\n data: JSON.stringify({ id: step.id, critique: critiqueResult }),\n });\n } catch (error) {\n console.error(`Step ${step.id} failed:`, error);\n planMap.get(step.id).status = \"error\";\n completedSteps.add(step.id); // Also add failed steps to avoid blocking\n await streamApi.writeSSE({\n event: \"step_error\",\n data: JSON.stringify({ id: step.id, error: error.message }),\n });\n }\n }));\n }\n await streamApi.writeSSE({\n event: \"project_complete\",\n data: \"Project finished\",\n });\n } catch (error) {\n console.error(\"Project flow failed:\", error);\n await streamApi.writeSSE({\n event: \"error\",\n data: `An error occurred: ${error.message}`,\n });\n }\n });\n});\n\n// --- Totem API Routes (Integrated into Hono) ---\napp.post(\"/api/totem/start\", async (c: Context) => {\n const { claim } = await c.req.json();\n if (!claim || typeof claim !== \"string\" || claim.trim().length === 0) {\n return c.json(\n { error: \"A valid, non-empty 'claim' string is required.\" },\n 400,\n );\n }\n const jobId = nanoid();\n const rootClaim: ClaimNode = {\n id: nanoid(),\n claim: claim,\n children: [],\n depth: 0,\n };\n await blob.setJSON(jobId, { status: \"processing\", totem: rootClaim });\n // This runs in the background and does not block the response\n buildTotemInBackground(jobId, rootClaim);\n return c.json({ jobId });\n});\n\napp.get(\"/api/totem/status\", async (c: Context) => {\n const jobId = c.req.query(\"jobId\");\n if (!jobId) {\n return c.json({ error: \"A 'jobId' query parameter is required.\" }, 400);\n }\n const state: JobState | undefined = await blob.getJSON(jobId);\n if (!state) {\n return c.json({ status: \"error\", error: \"Job not found.\" }, 404);\n }\n return c.json(state);\n});\n\napp.get(\"/api/totem/list\", async (c: Context) => {\n const index = await blob.getJSON(TOTEM_INDEX_KEY) || [];\n return c.json(index);\n});\n\napp.delete(\"/api/totem/delete\", async (c: Context) => {\n const jobId = c.req.query(\"jobId\");\n if (!jobId) {\n return c.json({ error: \"A 'jobId' query parameter is required.\" }, 400);\n }\n const currentIndex: TotemIndexEntry[] = await blob.getJSON(TOTEM_INDEX_KEY) ||\n [];\n const newIndex = currentIndex.filter((item) => item.jobId !== jobId);\n await blob.setJSON(TOTEM_INDEX_KEY, newIndex);\n await blob.delete(jobId);\n return c.json({ success: true, jobId });\n});\n\nexport default app.fetch;","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":14,"_58":59,"_18":60},["D",1760285775033],"children",[],[35],"a2c8f7a2-a779-11f0-91b4-0224a6c84d84",["D",1760280131240],"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":77,"_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":14,"_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":14,"_58":146,"_18":147},[],[35],["D",1760280131240],{"_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","fdb67791413f4386b4dbf18f354029a617f6160f","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"]