[{"_1":2,"_3":-5,"_4":-5},"loaderData",{"_5":6,"_7":8,"_9":10},"actionData","errors","root",{"_339":340,"_341":342},"routes/_app.x.$handle.$name",{"_97":96,"_32":45,"_106":87,"_27":107,"_13":14,"_108":-5,"_11":109,"_110":111,"_112":113,"_114":115,"_116":117,"_20":-5,"_118":-5,"_119":-5,"_120":25,"_121":122,"_123":59,"_124":-5,"_125":-5,"_126":127},"routes/_app.x.$handle.$name.branch.$branch.version.$version.diff",{"_11":12,"_13":14,"_15":16,"_17":-5,"_18":-5},"project",{"_43":44,"_32":45,"_38":46,"_47":-5,"_48":25,"_49":50,"_51":50,"_52":53,"_54":-5,"_18":55,"_56":57,"_58":59,"_60":-5,"_61":-5,"_62":63,"_64":65,"_66":-5,"_67":-5,"_68":-5,"_69":-5,"_70":71,"_72":71,"_73":74,"_75":74,"_76":74,"_77":71,"_78":71,"_79":80,"_81":82,"_83":84,"_85":25},"version",78,"changes",[19],"aiChatInfo","userId",{"_20":21,"_22":23,"_24":25,"_26":25,"_27":28,"_29":28},"file",{"_30":31,"_32":28,"_33":34,"_35":37,"_38":39,"_40":14,"_41":42,"_24":25},"previous",{"_30":31,"_32":28,"_33":34,"_35":36},"deleted",false,"deletedHere","path","verify.ts","latestPath","fileId","cae880f6-acf8-11f1-b03c-1607ee4eb77e","name","fileType","script","content","const BASE = \"https://rumbo-crm-openai-staging.val.run\";\nconst EXPECTED_SHA = \"9152f5a50de42be55508f0effed5ed09aee44c60\";\nconst PROTOCOL = \"2025-06-18\";\nconst expectedTools = [\n \"crm_get_service_readiness\",\n \"crm_prepare_note\",\n \"crm_prepare_lead_stage_change\",\n \"crm_prepare_outreach_draft\",\n];\nconst transport = {\n origin: \"https://chatgpt.com\",\n accept: \"application/json, text/event-stream\",\n \"content-type\": \"application/json\",\n};\nconst negotiated = { ...transport, \"mcp-protocol-version\": PROTOCOL };\nfunction assert(condition: unknown, message: string): asserts condition {\n if (!condition) throw new Error(message);\n}\nasync function expectStatus(name: string, response: Response, expected: number) {\n if (response.status !== expected) throw new Error(`${name}: expected ${expected}, got ${response.status}: ${(await response.text()).slice(0, 300)}`);\n console.log(`PASS ${name} ${expected}`);\n return response;\n}\nlet listedTools: Array = [];\nfunction assertResponseHygiene(toolName: string, output: Record) {\n const tool = listedTools.find((t) => t.name === toolName);\n assert(tool, `${toolName}: missing tool definition`);\n const allowed = new Set(Object.keys(tool.outputSchema?.properties ?? {}));\n const unexpected = Object.keys(output).filter((key) => !allowed.has(key));\n assert(unexpected.length === 0, `${toolName}: unexpected output fields ${unexpected.join(\",\")}`);\n const sensitiveKey = /(password|passwd|secret|api[_-]?key|access[_-]?token|refresh[_-]?token|authorization|cookie|stack|trace|debug)/i;\n const walk = (value: unknown, path = \"output\") => {\n if (Array.isArray(value)) return value.forEach((v, i) => walk(v, `${path}[${i}]`));\n if (value && typeof value === \"object\") {\n for (const [key, child] of Object.entries(value as Record)) {\n assert(!sensitiveKey.test(key), `${toolName}: sensitive/debug key ${path}.${key}`);\n walk(child, `${path}.${key}`);\n }\n return;\n }\n if (typeof value === \"string\") {\n assert(!/\\bsk-(?:proj-)?[A-Za-z0-9_-]{12,}\\b|\\bBearer\\s+[A-Za-z0-9._~-]{12,}/.test(value), `${toolName}: secret-like value at ${path}`);\n }\n };\n walk(output);\n}\nasync function rpc(id: number, name: string, args: Record) {\n const res = await fetch(`${BASE}/api/openai-plugin/mcp`, {\n method: \"POST\", headers: negotiated,\n body: JSON.stringify({ jsonrpc: \"2.0\", id, method: \"tools/call\", params: { name, arguments: args } }),\n });\n await expectStatus(name, res, 200);\n const body = await res.json();\n assert(body?.result?.isError === false, `${name}: isError`);\n const output = body.result.structuredContent as Record;\n assertResponseHygiene(name, output);\n return output;\n}\n\nawait expectStatus(\"domain challenge unset\", await fetch(`${BASE}/.well-known/openai-apps-challenge`), 404);\nconst health = await expectStatus(\"health\", await fetch(`${BASE}/api/openai-plugin/health`, { headers: { origin: \"https://chatgpt.com\", accept: \"application/json\" } }), 200);\nconst hb = await health.json();\nassert(hb.sourceSha === EXPECTED_SHA, \"source SHA mismatch\");\nassert(hb.customerDataToolsExposed === false, \"customer data unexpectedly exposed\");\nassert(hb.productionAuthority === false && hb.submissionAuthority === false, \"authority escalation\");\n\nawait expectStatus(\"GET without SSE\", await fetch(`${BASE}/api/openai-plugin/mcp`, { method: \"GET\", headers: { origin: \"https://chatgpt.com\" } }), 405);\nawait expectStatus(\"Origin deny\", await fetch(`${BASE}/api/openai-plugin/mcp`, { method: \"POST\", headers: { ...transport, origin: \"https://evil.example\" }, body: \"{}\" }), 403);\nawait expectStatus(\"Accept deny\", await fetch(`${BASE}/api/openai-plugin/mcp`, { method: \"POST\", headers: { ...transport, accept: \"application/json\" }, body: \"{}\" }), 406);\nawait expectStatus(\"Content-Type deny\", await fetch(`${BASE}/api/openai-plugin/mcp`, { method: \"POST\", headers: { ...transport, \"content-type\": \"text/plain\" }, body: \"{}\" }), 415);\n\nconst init = await expectStatus(\"initialize\", await fetch(`${BASE}/api/openai-plugin/mcp`, {\n method: \"POST\", headers: transport,\n body: JSON.stringify({ jsonrpc: \"2.0\", id: 1, method: \"initialize\", params: { protocolVersion: PROTOCOL, capabilities: {}, clientInfo: { name: \"rumbo-valtown-verifier\", version: \"1.0.0\" } } }),\n}), 200);\nconst initBody = await init.json();\nassert(initBody?.result?.protocolVersion === PROTOCOL, \"initialize protocol mismatch\");\n\nawait expectStatus(\"missing protocol deny\", await fetch(`${BASE}/api/openai-plugin/mcp`, {\n method: \"POST\", headers: transport,\n body: JSON.stringify({ jsonrpc: \"2.0\", id: 2, method: \"tools/list\", params: {} }),\n}), 400);\n\nconst list = await expectStatus(\"tools/list\", await fetch(`${BASE}/api/openai-plugin/mcp`, {\n method: \"POST\", headers: negotiated,\n body: JSON.stringify({ jsonrpc: \"2.0\", id: 3, method: \"tools/list\", params: {} }),\n}), 200);\nconst listBody = await list.json();\nconst names = listBody?.result?.tools?.map((tool: { name: string }) => tool.name) ?? [];\nassert(JSON.stringify(names) === JSON.stringify(expectedTools), `tool catalog mismatch: ${JSON.stringify(names)}`);\nfor (const tool of listBody.result.tools) {\n assert(tool.annotations?.readOnlyHint === true, `${tool.name}: readOnlyHint`);\n assert(tool.annotations?.destructiveHint === false, `${tool.name}: destructiveHint`);\n assert(tool.annotations?.openWorldHint === false, `${tool.name}: openWorldHint`);\n assert(tool.inputSchema?.additionalProperties === false, `${tool.name}: open input schema`);\n assert(tool.outputSchema?.additionalProperties === false, `${tool.name}: open output schema`);\n}\n\nconst initialized = await expectStatus(\"notifications/initialized\", await fetch(`${BASE}/api/openai-plugin/mcp`, {\n method: \"POST\", headers: negotiated,\n body: JSON.stringify({ jsonrpc: \"2.0\", method: \"notifications/initialized\", params: {} }),\n}), 202);\nassert((await initialized.text()).length === 0, \"notification body must be empty\");\n\nconst readiness = await rpc(10, \"crm_get_service_readiness\", {});\nassert(readiness.customer_data_tools_exposed === false && readiness.mutating_tools_exposed === false, \"readiness exposure\");\nassert(readiness.production_authority === false && readiness.submission_authority === false, \"readiness authority\");\n\nconst note = await rpc(11, \"crm_prepare_note\", { workspace_id: \"ws-demo\", subject_id: \"lead-demo\", note: \"Synthetic conformance test\" });\nassert(note.kind === \"crm_note_proposal\" && note.state === \"REVIEW_REQUIRED\" && note.side_effect === false, \"note proposal contract\");\n\nconst stage = await rpc(12, \"crm_prepare_lead_stage_change\", { workspace_id: \"ws-demo\", lead_id: \"lead-demo\", from_stage: \"new\", to_stage: \"qualified\", reason: \"Synthetic conformance test\" });\nassert(stage.kind === \"lead_stage_change_proposal\" && stage.state === \"REVIEW_REQUIRED\" && stage.side_effect === false, \"stage proposal contract\");\n\nconst outreach = await rpc(13, \"crm_prepare_outreach_draft\", { workspace_id: \"ws-demo\", lead_id: \"lead-demo\", channel: \"whatsapp\", objective: \"Synthetic conformance test\" });\nassert(outreach.kind === \"outreach_draft_request\" && outreach.state === \"REVIEW_REQUIRED\" && outreach.side_effect === false && outreach.send_authorized === false, \"outreach proposal contract\");\n\nconsole.log(`PASS source SHA ${EXPECTED_SHA}`);\nconsole.log(\"RUMBO_CRM_VALTOWN_STAGING_CONFORMANCE_PASS\");\n","const BASE = \"https://rumbo-crm-openai-staging.val.run\";\nconst EXPECTED_SHA = \"9152f5a50de42be55508f0effed5ed09aee44c60\";\nconst PROTOCOL = \"2025-06-18\";\nconst expectedTools = [\n \"crm_get_service_readiness\",\n \"crm_prepare_note\",\n \"crm_prepare_lead_stage_change\",\n \"crm_prepare_outreach_draft\",\n];\nconst transport = {\n origin: \"https://chatgpt.com\",\n accept: \"application/json, text/event-stream\",\n \"content-type\": \"application/json\",\n};\nconst negotiated = { ...transport, \"mcp-protocol-version\": PROTOCOL };\nfunction assert(condition: unknown, message: string): asserts condition {\n if (!condition) throw new Error(message);\n}\nasync function expectStatus(name: string, response: Response, expected: number) {\n if (response.status !== expected) throw new Error(`${name}: expected ${expected}, got ${response.status}: ${(await response.text()).slice(0, 300)}`);\n console.log(`PASS ${name} ${expected}`);\n return response;\n}\nlet listedTools: Array = [];\nfunction assertResponseHygiene(toolName: string, output: Record) {\n const tool = listedTools.find((t) => t.name === toolName);\n assert(tool, `${toolName}: missing tool definition`);\n const allowed = new Set(Object.keys(tool.outputSchema?.properties ?? {}));\n const unexpected = Object.keys(output).filter((key) => !allowed.has(key));\n assert(unexpected.length === 0, `${toolName}: unexpected output fields ${unexpected.join(\",\")}`);\n const sensitiveKey = /(password|passwd|secret|api[_-]?key|access[_-]?token|refresh[_-]?token|authorization|cookie|stack|trace|debug)/i;\n const walk = (value: unknown, path = \"output\") => {\n if (Array.isArray(value)) return value.forEach((v, i) => walk(v, `${path}[${i}]`));\n if (value && typeof value === \"object\") {\n for (const [key, child] of Object.entries(value as Record)) {\n assert(!sensitiveKey.test(key), `${toolName}: sensitive/debug key ${path}.${key}`);\n walk(child, `${path}.${key}`);\n }\n return;\n }\n if (typeof value === \"string\") {\n assert(!/\\bsk-(?:proj-)?[A-Za-z0-9_-]{12,}\\b|\\bBearer\\s+[A-Za-z0-9._~-]{12,}/.test(value), `${toolName}: secret-like value at ${path}`);\n }\n };\n walk(output);\n}\nasync function rpc(id: number, name: string, args: Record) {\n const res = await fetch(`${BASE}/api/openai-plugin/mcp`, {\n method: \"POST\", headers: negotiated,\n body: JSON.stringify({ jsonrpc: \"2.0\", id, method: \"tools/call\", params: { name, arguments: args } }),\n });\n await expectStatus(name, res, 200);\n const body = await res.json();\n assert(body?.result?.isError === false, `${name}: isError`);\n const output = body.result.structuredContent as Record;\n assertResponseHygiene(name, output);\n return output;\n}\n\nawait expectStatus(\"domain challenge unset\", await fetch(`${BASE}/.well-known/openai-apps-challenge`), 404);\nconst health = await expectStatus(\"health\", await fetch(`${BASE}/api/openai-plugin/health`, { headers: { origin: \"https://chatgpt.com\", accept: \"application/json\" } }), 200);\nconst hb = await health.json();\nassert(hb.sourceSha === EXPECTED_SHA, \"source SHA mismatch\");\nassert(hb.customerDataToolsExposed === false, \"customer data unexpectedly exposed\");\nassert(hb.productionAuthority === false && hb.submissionAuthority === false, \"authority escalation\");\n\nawait expectStatus(\"GET without SSE\", await fetch(`${BASE}/api/openai-plugin/mcp`, { method: \"GET\", headers: { origin: \"https://chatgpt.com\" } }), 405);\nawait expectStatus(\"Origin deny\", await fetch(`${BASE}/api/openai-plugin/mcp`, { method: \"POST\", headers: { ...transport, origin: \"https://evil.example\" }, body: \"{}\" }), 403);\nawait expectStatus(\"Accept deny\", await fetch(`${BASE}/api/openai-plugin/mcp`, { method: \"POST\", headers: { ...transport, accept: \"application/json\" }, body: \"{}\" }), 406);\nawait expectStatus(\"Content-Type deny\", await fetch(`${BASE}/api/openai-plugin/mcp`, { method: \"POST\", headers: { ...transport, \"content-type\": \"text/plain\" }, body: \"{}\" }), 415);\n\nconst init = await expectStatus(\"initialize\", await fetch(`${BASE}/api/openai-plugin/mcp`, {\n method: \"POST\", headers: transport,\n body: JSON.stringify({ jsonrpc: \"2.0\", id: 1, method: \"initialize\", params: { protocolVersion: PROTOCOL, capabilities: {}, clientInfo: { name: \"rumbo-valtown-verifier\", version: \"1.0.0\" } } }),\n}), 200);\nconst initBody = await init.json();\nassert(initBody?.result?.protocolVersion === PROTOCOL, \"initialize protocol mismatch\");\n\nawait expectStatus(\"missing protocol deny\", await fetch(`${BASE}/api/openai-plugin/mcp`, {\n method: \"POST\", headers: transport,\n body: JSON.stringify({ jsonrpc: \"2.0\", id: 2, method: \"tools/list\", params: {} }),\n}), 400);\n\nconst list = await expectStatus(\"tools/list\", await fetch(`${BASE}/api/openai-plugin/mcp`, {\n method: \"POST\", headers: negotiated,\n body: JSON.stringify({ jsonrpc: \"2.0\", id: 3, method: \"tools/list\", params: {} }),\n}), 200);\nconst listBody = await list.json();\nlistedTools = listBody?.result?.tools ?? [];\nconst names = listedTools.map((tool: { name: string }) => tool.name);\nassert(JSON.stringify(names) === JSON.stringify(expectedTools), `tool catalog mismatch: ${JSON.stringify(names)}`);\nfor (const tool of listedTools) {\n assert(tool.annotations?.readOnlyHint === true, `${tool.name}: readOnlyHint`);\n assert(tool.annotations?.destructiveHint === false, `${tool.name}: destructiveHint`);\n assert(tool.annotations?.openWorldHint === false, `${tool.name}: openWorldHint`);\n assert(tool.inputSchema?.additionalProperties === false, `${tool.name}: open input schema`);\n assert(tool.outputSchema?.additionalProperties === false, `${tool.name}: open output schema`);\n}\n\nconst initialized = await expectStatus(\"notifications/initialized\", await fetch(`${BASE}/api/openai-plugin/mcp`, {\n method: \"POST\", headers: negotiated,\n body: JSON.stringify({ jsonrpc: \"2.0\", method: \"notifications/initialized\", params: {} }),\n}), 202);\nassert((await initialized.text()).length === 0, \"notification body must be empty\");\n\nconst readiness = await rpc(10, \"crm_get_service_readiness\", {});\nassert(readiness.customer_data_tools_exposed === false && readiness.mutating_tools_exposed === false, \"readiness exposure\");\nassert(readiness.production_authority === false && readiness.submission_authority === false, \"readiness authority\");\n\nconst note = await rpc(11, \"crm_prepare_note\", { workspace_id: \"ws-demo\", subject_id: \"lead-demo\", note: \"Synthetic conformance test\" });\nassert(note.kind === \"crm_note_proposal\" && note.state === \"REVIEW_REQUIRED\" && note.side_effect === false, \"note proposal contract\");\n\nconst stage = await rpc(12, \"crm_prepare_lead_stage_change\", { workspace_id: \"ws-demo\", lead_id: \"lead-demo\", from_stage: \"new\", to_stage: \"qualified\", reason: \"Synthetic conformance test\" });\nassert(stage.kind === \"lead_stage_change_proposal\" && stage.state === \"REVIEW_REQUIRED\" && stage.side_effect === false, \"stage proposal contract\");\n\nconst outreach = await rpc(13, \"crm_prepare_outreach_draft\", { workspace_id: \"ws-demo\", lead_id: \"lead-demo\", channel: \"whatsapp\", objective: \"Synthetic conformance test\" });\nassert(outreach.kind === \"outreach_draft_request\" && outreach.state === \"REVIEW_REQUIRED\" && outreach.side_effect === false && outreach.send_authorized === false, \"outreach proposal contract\");\n\nconsole.log(`PASS source SHA ${EXPECTED_SHA}`);\nconsole.log(\"RUMBO_CRM_VALTOWN_STAGING_CONFORMANCE_PASS\");\n","createdAt",["D",1789115828327],"branchVersion","branchId","ceba545d-7cc1-405d-a0c7-d27e7e735161","id","3255f42e-7bba-4583-8764-631437397e4e","rumbo-crm-openai-staging",["D",1789031650837],"deletedAt","legacyEnvVars","privacy","public","httpPrivacy","description","RUMBO IA CRM bounded MCP staging","imageUrl","d3788f19-13c5-4356-87e8-dce639626bb9","pinnedFileId","fbed6f64-acf7-11f1-b3b4-1607ee4eb77e","pinPreview",true,"pinMigrated","screenshotVersion","likeCount",0,"orgId","01a04116-f28c-754c-9173-77bf5814fcd6","tursoName","tursoHostname","tursoMeasuredAt","pinned","userPerm","r","valPerm","blobPerm","rw","emailPerm","sqlitePerm","projectPerm","telemetryPerm","user",{"_97":96,"_105":101},"org",{"_97":96,"_98":99,"_100":101,"_18":55,"_102":103,"_104":25},"branch",{"_13":86,"_32":87,"_43":42,"_88":44,"_89":-5,"_90":-5,"_91":-5,"_92":-5,"_93":-5,"_94":45,"_95":96},"isOwned",284,"main","projectId","forkedBranchId","forkedBranchVersion","forkedProjectPrivacy","forkedProjectId","forkedProjectOrgId","projectName","userHandle","sebas1","handle","displayName","sebas","avatar","https://lh3.googleusercontent.com/a/ACg8ocJaC6uiXPmN9dvgA2XcSlzBSkPJO0GN5zKzPm3a8cbjmXC9HYA=s96-c","tier","free","canRestrictVals","profileImageUrl","branchName","","adoptOrgId",{"_43":44,"_32":45,"_38":335,"_47":-5,"_48":25,"_49":50,"_51":50,"_52":53,"_54":-5,"_18":55,"_56":57,"_58":59,"_60":-5,"_61":-5,"_62":63,"_64":65,"_66":-5,"_67":-5,"_68":-5,"_69":-5,"_70":71,"_72":71,"_73":74,"_75":74,"_76":74,"_77":71,"_78":71,"_79":336,"_81":337,"_83":338,"_85":25},"tree",[270,271,272,273,274,275,276,277,278,279,280,281,282,283,284,285],"files",[270,271,272,273,274,275,276,277,278,279,280,281,282,283,284,285],"counts",{"_260":63,"_261":243,"_262":63,"_263":264,"_265":266},"contents",[128,129,130,131,132,133,134,135,136,137,138,139,140,141,142,143],"upstreamProject","upstreamBranch","emptyEnvVars","flags",[],"hasHttpFiles","grants","orgs","customDomains",{},{"_43":253,"_32":254,"_30":255,"_147":-5,"_41":42,"_40":256,"_149":59,"_38":257,"_13":63,"_33":20,"_35":258,"_24":25,"_152":-5,"_27":259},{"_43":246,"_32":247,"_30":248,"_147":-5,"_41":42,"_40":249,"_149":25,"_38":250,"_13":207,"_33":20,"_35":251,"_24":25,"_152":-5,"_27":252},{"_43":238,"_32":239,"_30":240,"_147":-5,"_41":42,"_40":241,"_149":25,"_38":242,"_13":243,"_33":20,"_35":244,"_24":25,"_152":-5,"_27":245},{"_43":231,"_32":232,"_30":233,"_147":-5,"_41":42,"_40":234,"_149":59,"_38":235,"_13":63,"_33":20,"_35":236,"_24":25,"_152":-5,"_27":237},{"_43":224,"_32":225,"_30":226,"_147":-5,"_41":42,"_40":227,"_149":25,"_38":228,"_13":163,"_33":20,"_35":229,"_24":25,"_152":-5,"_27":230},{"_43":217,"_32":218,"_30":219,"_147":-5,"_41":42,"_40":220,"_149":59,"_38":221,"_13":63,"_33":20,"_35":222,"_24":25,"_152":-5,"_27":223},{"_43":210,"_32":211,"_30":212,"_147":-5,"_41":42,"_40":213,"_149":59,"_38":214,"_13":63,"_33":20,"_35":215,"_24":25,"_152":-5,"_27":216},{"_43":202,"_32":203,"_30":204,"_147":-5,"_41":42,"_40":205,"_149":59,"_38":206,"_13":207,"_33":20,"_35":208,"_24":25,"_152":-5,"_27":209},{"_43":195,"_32":196,"_30":197,"_147":-5,"_41":42,"_40":198,"_149":25,"_38":199,"_13":184,"_33":20,"_35":200,"_24":25,"_152":-5,"_27":201},{"_43":188,"_32":189,"_30":190,"_147":-5,"_41":42,"_40":191,"_149":25,"_38":192,"_13":63,"_33":20,"_35":193,"_24":25,"_152":-5,"_27":194},{"_43":181,"_32":182,"_30":183,"_147":-5,"_41":42,"_40":184,"_149":59,"_38":185,"_13":63,"_33":20,"_35":186,"_24":25,"_152":-5,"_27":187},{"_43":174,"_32":175,"_30":176,"_147":-5,"_41":42,"_40":177,"_149":59,"_38":178,"_13":63,"_33":20,"_35":179,"_24":25,"_152":-5,"_27":180},{"_43":166,"_32":167,"_30":57,"_147":-5,"_41":42,"_40":168,"_149":25,"_38":169,"_13":170,"_33":171,"_35":172,"_24":25,"_152":45,"_27":173},{"_43":158,"_32":159,"_30":160,"_147":-5,"_41":42,"_40":161,"_149":25,"_38":162,"_13":163,"_33":34,"_35":164,"_24":25,"_152":-5,"_27":165},{"_43":154,"_32":28,"_30":31,"_147":-5,"_41":42,"_40":14,"_149":25,"_38":155,"_13":156,"_33":34,"_35":37,"_24":25,"_152":-5,"_27":157},{"_43":144,"_32":145,"_30":146,"_147":-5,"_41":42,"_40":148,"_149":59,"_38":150,"_13":63,"_33":34,"_35":151,"_24":25,"_152":-5,"_27":153},"6a5ab433-b141-45f6-8966-4592b1bb3580","verify_submission_import.ts","8fe4ec18-ada4-11f1-9cb0-1607ee4eb77e","parentId",61,"latest",["D",1789105772539],"import submission from \"./chatgpt-app-submission.json\" with { type: \"json\" };\n\nconst EXPECTED_TOOLS = [\n \"crm_get_service_readiness\",\n \"crm_prepare_note\",\n \"crm_prepare_lead_stage_change\",\n \"crm_prepare_outreach_draft\",\n] as const;\nconst ALLOWED_CATEGORIES = new Set([\"BUSINESS\",\"COLLABORATION\",\"DESIGN\",\"DEVELOPER_TOOLS\",\"EDUCATION\",\"ENTERTAINMENT\",\"FINANCE\",\"FOOD\",\"LIFESTYLE\",\"NEWS\",\"PRODUCTIVITY\",\"SHOPPING\",\"TRAVEL\"]);\nconst failures: string[] = [];\nconst pass = (name: string, ok: boolean) => { console.log(`${ok ? \"PASS\" : \"FAIL\"} ${name}`); if (!ok) failures.push(name); };\n\npass(\"schema_version=1\", submission.schema_version === 1);\npass(\"schema_url\", submission.$schema === \"https://developers.openai.com/apps-sdk/schemas/chatgpt-app-submission.v1.json\");\npass(\"display_name\", submission.app_info?.display_name === \"RUMBO IA CRM\");\npass(\"subtitle<=30\", typeof submission.app_info?.subtitle === \"string\" && submission.app_info.subtitle.length <= 30);\npass(\"category\", ALLOWED_CATEGORIES.has(submission.app_info?.category));\n\nconst toolNames = Object.keys(submission.tools ?? {}).sort();\npass(\"exact_four_tools\", JSON.stringify(toolNames) === JSON.stringify([...EXPECTED_TOOLS].sort()));\nfor (const name of EXPECTED_TOOLS) {\n const t = submission.tools?.[name];\n pass(`${name}.annotations`, t?.annotations?.readOnlyHint === true && t?.annotations?.openWorldHint === false && t?.annotations?.destructiveHint === false);\n pass(`${name}.justifications`, [t?.justifications?.read_only_justification, t?.justifications?.open_world_justification, t?.justifications?.destructive_justification].every((v) => typeof v === \"string\" && v.trim().length > 0));\n}\n\npass(\"five_positive_tests\", Array.isArray(submission.test_cases) && submission.test_cases.length === 5);\npass(\"three_negative_tests\", Array.isArray(submission.negative_test_cases) && submission.negative_test_cases.length === 3);\nfor (const [i, tc] of (submission.test_cases ?? []).entries()) {\n pass(`positive_${i + 1}_tool_valid`, EXPECTED_TOOLS.includes(tc.tools_triggered));\n pass(`positive_${i + 1}_fields`, [tc.description, tc.user_prompt, tc.expected_output].every((v) => typeof v === \"string\" && v.trim().length > 0));\n}\nfor (const [i, tc] of (submission.negative_test_cases ?? []).entries()) {\n pass(`negative_${i + 1}_no_tool`, tc.tools_triggered === null);\n pass(`negative_${i + 1}_fields`, [tc.description, tc.user_prompt, tc.expected_output].every((v) => typeof v === \"string\" && v.trim().length > 0));\n}\n\nif (failures.length) {\n console.log(JSON.stringify({ state: \"FAIL\", failures }, null, 2));\n throw new Error(`Submission import verifier failed: ${failures.join(\", \")}`);\n}\nconsole.log(JSON.stringify({ state: \"PASS\", tools: EXPECTED_TOOLS.length, positive_tests: submission.test_cases.length, negative_tests: submission.negative_test_cases.length, money_spend: 0 }, null, 2));\nconsole.log(\"RUMBO_CHATGPT_APP_SUBMISSION_IMPORT_PASS\");\n","label",[145],"7d9777c9-3576-49f8-84bf-01ada76a4772",["D",1789115828327],4,[28],"e467aec1-b612-4170-bb7d-448f7259a04e","submission_gate.ts","fdf0ffb8-ad58-11f1-888a-1607ee4eb77e",70,["D",1789110934250],16,"const BASE = \"https://rumbo-crm-openai-staging.val.run\";\nconst EXPECTED_SHA = \"9152f5a50de42be55508f0effed5ed09aee44c60\";\nconst EXPECTED_LOGO_SHA256 = \"3ac48eb17d735ee41d12e6edf0cc0ea0e5a55358cce6b5fd37d93b3fc0372205\";\nconst LISTING_URLS = {\n website: `${BASE}/`,\n support: `${BASE}/support`,\n privacy: `${BASE}/privacy`,\n terms: `${BASE}/terms`,\n};\n\nconst blockers: string[] = [];\nconst facts: Record = {};\n\nasync function check(name: string, fn: () => Promise) {\n try {\n const ok = await fn();\n facts[name] = ok ? \"PASS\" : \"FAIL\";\n if (!ok) blockers.push(name);\n } catch (e) {\n facts[name] = `ERROR:${e instanceof Error ? e.message : String(e)}`;\n blockers.push(name);\n }\n}\nasync function sha256Hex(bytes: Uint8Array) {\n const digest = new Uint8Array(await crypto.subtle.digest(\"SHA-256\", bytes));\n return Array.from(digest, (b) => b.toString(16).padStart(2, \"0\")).join(\"\");\n}\n\nawait check(\"health_exact_sha\", async () => {\n const r = await fetch(`${BASE}/api/openai-plugin/health`, { headers: { origin: \"https://chatgpt.com\", accept: \"application/json\" } });\n if (r.status !== 200) return false;\n const b = await r.json();\n return b?.sourceSha === EXPECTED_SHA && b?.productionAuthority === false && b?.submissionAuthority === false;\n});\n\nawait check(\"tool_annotations\", async () => {\n const headers = { origin: \"https://chatgpt.com\", accept: \"application/json, text/event-stream\", \"content-type\": \"application/json\", \"mcp-protocol-version\": \"2025-06-18\" };\n const r = await fetch(`${BASE}/api/openai-plugin/mcp`, { method: \"POST\", headers, body: JSON.stringify({ jsonrpc: \"2.0\", id: 1, method: \"tools/list\", params: {} }) });\n if (r.status !== 200) return false;\n const b = await r.json();\n const tools = b?.result?.tools ?? [];\n return tools.length === 4 && tools.every((t: any) => t.annotations?.readOnlyHint === true && t.annotations?.openWorldHint === false && t.annotations?.destructiveHint === false);\n});\n\nawait check(\"listing_urls_public\", async () => {\n const responses = await Promise.all(Object.values(LISTING_URLS).map((url) => fetch(url, { redirect: \"follow\" })));\n facts.listing_url_statuses = Object.fromEntries(Object.keys(LISTING_URLS).map((key, i) => [key, responses[i].status]));\n return responses.every((r) => r.status >= 200 && r.status < 300);\n});\n\nawait check(\"support_mcp_semantics\", async () => {\n const r = await fetch(LISTING_URLS.support, { redirect: \"follow\" });\n if (!r.ok) return false;\n const text = (await r.text()).toLowerCase();\n const namesCrm = text.includes(\"rumbo crm\") || text.includes(\"rumbo ia crm\");\n const describesBoundedSupport = text.includes(\"proposal-only\") || text.includes(\"review\") || text.includes(\"proposal tools\");\n const namesOpenAiSurface = text.includes(\"openai\") || text.includes(\"chatgpt\");\n const hasSupportContact = text.includes(\"sebastian@rumbo.verso.fans\") || text.includes(\"contact support\");\n return namesCrm && describesBoundedSupport && namesOpenAiSurface && hasSupportContact;\n});\n\nawait check(\"privacy_mcp_disclosure\", async () => {\n const r = await fetch(LISTING_URLS.privacy, { redirect: \"follow\" });\n if (!r.ok) return false;\n const text = (await r.text()).toLowerCase();\n const namesMcpSurface = text.includes(\"mcp\") || text.includes(\"chatgpt\") || text.includes(\"plugin\");\n const describesSubmittedContent = text.includes(\"note\") || text.includes(\"lead\") || text.includes(\"workspace\") || text.includes(\"identifier\");\n return namesMcpSurface && describesSubmittedContent;\n});\n\nawait check(\"terms_bounded_semantics\", async () => {\n const r = await fetch(LISTING_URLS.terms, { redirect: \"follow\" });\n if (!r.ok) return false;\n const text = (await r.text()).toLowerCase();\n return text.includes(\"human-controlled\") && text.includes(\"proposals\") && text.includes(\"not automatically applied\");\n});\n\nawait check(\"logo_exact_asset\", async () => {\n const r = await fetch(`${BASE}/logo.svg`, { redirect: \"follow\" });\n if (!r.ok || !(r.headers.get(\"content-type\") ?? \"\").includes(\"image/svg+xml\")) return false;\n const hash = await sha256Hex(new Uint8Array(await r.arrayBuffer()));\n facts.logo_sha256 = hash;\n return hash === EXPECTED_LOGO_SHA256;\n});\n\nconst challenge = await fetch(`${BASE}/.well-known/openai-apps-challenge`);\nfacts.domain_challenge_status = challenge.status;\nif (challenge.status !== 200) blockers.push(\"domain_verification_not_completed\");\n\nblockers.push(\"portal_app_creation_or_submission_access_unverified\");\nblockers.push(\"publisher_identity_portal_requirement_unresolved\");\nfacts.project_data_residency = \"UNVERIFIED_NOT_CURRENT_OFFICIAL_SUBMISSION_BLOCKER\";\nfacts.logo_asset = \"CANONICAL_REMOTE_MAIN_VERIFIED_SHA256_3ac48eb17d735ee41d12e6edf0cc0ea0e5a55358cce6b5fd37d93b3fc0372205\";\nblockers.push(\"logo_portal_upload_not_executed\");\nblockers.push(\"scan_tools_not_executed\");\nblockers.push(\"distribution_scope_unset\");\nblockers.push(\"production_mcp_url_not_authorized\");\n\nconst result = {\n state: blockers.length === 0 ? \"SUBMISSION_READY\" : \"BLOCKED_FAIL_CLOSED\",\n candidate_sha: EXPECTED_SHA,\n money_spend: 0,\n facts,\n blockers,\n authority: \"NOT_GRANTED_BY_TECHNICAL_GATE\",\n};\n\nconsole.log(JSON.stringify(result, null, 2));\nif (blockers.length === 0) console.log(\"RUMBO_OPENAI_SUBMISSION_GATE_PASS\");\nelse console.log(\"RUMBO_OPENAI_SUBMISSION_GATE_BLOCKED\");",[159],"cc34b718-a9c2-4bd2-a399-ee0b08368cc4","main.ts",51,["D",1789105438225],8,"http","const SOURCE_SHA = \"9152f5a50de42be55508f0effed5ed09aee44c60\";\nconst PROTOCOL_VERSION = \"2025-06-18\";\nconst FALLBACK_PROTOCOL_VERSION = \"2025-03-26\";\nconst OPENAI_APPS_CHALLENGE_TOKEN = Deno.env.get(\"OPENAI_APPS_CHALLENGE_TOKEN\")?.trim() ?? \"\";\nconst PORTAL_LOGO_SVG = `\n`;\nconst ALLOWED_ORIGINS = new Set([\n \"https://chatgpt.com\",\n \"https://chat.openai.com\",\n \"https://platform.openai.com\",\n]);\nconst EXPOSED_TOOLS = [\n \"crm_get_service_readiness\",\n \"crm_prepare_note\",\n \"crm_prepare_lead_stage_change\",\n \"crm_prepare_outreach_draft\",\n] as const;\nconst PLANNED_UNEXPOSED_TOOLS = [\n \"crm_search_contacts\",\n \"crm_get_contact\",\n \"crm_search_leads\",\n \"crm_get_pipeline_summary\",\n \"crm_search_conversations\",\n \"crm_search_knowledge\",\n \"crm_get_analytics_summary\",\n] as const;\n\nconst boundedId = { type: \"string\", minLength: 1, maxLength: 200, pattern: \"\\\\S\" };\nconst boundedString = (maxLength: number) => ({ type: \"string\", minLength: 1, maxLength, pattern: \"\\\\S\" });\nconst reviewBase = {\n state: { type: \"string\", const: \"REVIEW_REQUIRED\" },\n side_effect: { type: \"boolean\", const: false },\n};\nconst annotations = { readOnlyHint: true, destructiveHint: false, openWorldHint: false };\n\nconst tools = [\n {\n name: \"crm_get_service_readiness\",\n title: \"Get CRM service readiness\",\n description: \"Returns bounded publication/runtime readiness metadata. It does not access customer records.\",\n inputSchema: { type: \"object\", properties: {}, additionalProperties: false },\n outputSchema: {\n type: \"object\",\n properties: {\n state: { type: \"string\", const: \"MCP_CONTRACT_ACTIVE_BOUNDED\" },\n customer_data_tools_exposed: { type: \"boolean\", const: false },\n mutating_tools_exposed: { type: \"boolean\", const: false },\n production_authority: { type: \"boolean\", const: false },\n submission_authority: { type: \"boolean\", const: false },\n },\n required: [\"state\", \"customer_data_tools_exposed\", \"mutating_tools_exposed\", \"production_authority\", \"submission_authority\"],\n additionalProperties: false,\n },\n annotations,\n },\n {\n name: \"crm_prepare_note\",\n title: \"Prepare CRM note\",\n description: \"Creates a reviewable note proposal only. It does not write the note to the CRM.\",\n inputSchema: { type: \"object\", properties: { workspace_id: boundedId, subject_id: boundedId, note: boundedString(4000) }, required: [\"workspace_id\", \"subject_id\", \"note\"], additionalProperties: false },\n outputSchema: { type: \"object\", properties: { kind: { type: \"string\", const: \"crm_note_proposal\" }, ...reviewBase, workspace_id: { type: \"string\" }, subject_id: { type: \"string\" }, note: { type: \"string\" } }, required: [\"kind\", \"state\", \"side_effect\", \"workspace_id\", \"subject_id\", \"note\"], additionalProperties: false },\n annotations,\n },\n {\n name: \"crm_prepare_lead_stage_change\",\n title: \"Prepare lead stage change\",\n description: \"Creates a reviewable stage-change proposal only. It performs no CRM mutation.\",\n inputSchema: { type: \"object\", properties: { workspace_id: boundedId, lead_id: boundedId, from_stage: boundedId, to_stage: boundedId, reason: { type: \"string\", maxLength: 1000 } }, required: [\"workspace_id\", \"lead_id\", \"from_stage\", \"to_stage\"], additionalProperties: false },\n outputSchema: { type: \"object\", properties: { kind: { type: \"string\", const: \"lead_stage_change_proposal\" }, ...reviewBase, workspace_id: { type: \"string\" }, lead_id: { type: \"string\" }, from_stage: { type: \"string\" }, to_stage: { type: \"string\" }, reason: { type: \"string\" } }, required: [\"kind\", \"state\", \"side_effect\", \"workspace_id\", \"lead_id\", \"from_stage\", \"to_stage\", \"reason\"], additionalProperties: false },\n annotations,\n },\n {\n name: \"crm_prepare_outreach_draft\",\n title: \"Prepare outreach draft\",\n description: \"Returns a reviewable outreach draft request artifact. It never sends a message.\",\n inputSchema: { type: \"object\", properties: { workspace_id: boundedId, lead_id: boundedId, channel: { type: \"string\", enum: [\"email\", \"sms\", \"whatsapp\", \"other\"] }, objective: boundedString(2000) }, required: [\"workspace_id\", \"lead_id\", \"channel\", \"objective\"], additionalProperties: false },\n outputSchema: { type: \"object\", properties: { kind: { type: \"string\", const: \"outreach_draft_request\" }, ...reviewBase, workspace_id: { type: \"string\" }, lead_id: { type: \"string\" }, channel: { type: \"string\", enum: [\"email\", \"sms\", \"whatsapp\", \"other\"] }, objective: { type: \"string\" }, send_authorized: { type: \"boolean\", const: false } }, required: [\"kind\", \"state\", \"side_effect\", \"workspace_id\", \"lead_id\", \"channel\", \"objective\", \"send_authorized\"], additionalProperties: false },\n annotations,\n },\n];\n\ntype RecordLike = Record;\nfunction asRecord(value: unknown): RecordLike | null { return value !== null && typeof value === \"object\" && !Array.isArray(value) ? value as RecordLike : null; }\nfunction rpcOk(id: unknown, result: unknown) { return { jsonrpc: \"2.0\", id: id ?? null, result }; }\nfunction rpcError(id: unknown, code: number, message: string, data?: unknown) { return { jsonrpc: \"2.0\", id: id ?? null, error: { code, message, ...(data === undefined ? {} : { data }) } }; }\nfunction exactArgs(value: unknown, allowed: string[]) {\n if (value === undefined) return {} as RecordLike;\n const args = asRecord(value);\n if (!args) throw new Error(\"invalid_arguments_object\");\n const allow = new Set(allowed);\n if (Object.keys(args).some((key) => !allow.has(key))) throw new Error(\"undeclared_argument\");\n return args;\n}\nfunction requiredString(args: RecordLike, key: string, max: number) {\n const value = args[key];\n if (typeof value !== \"string\" || !value.trim() || value.length > max) throw new Error(`invalid_${key}`);\n return value.trim();\n}\nfunction optionalString(args: RecordLike, key: string, max: number) {\n const value = args[key];\n if (value === undefined) return \"\";\n if (typeof value !== \"string\" || value.length > max) throw new Error(`invalid_${key}`);\n return value.trim();\n}\nfunction proposal(kind: string, body: RecordLike) {\n const structuredContent = { kind, state: \"REVIEW_REQUIRED\", side_effect: false, ...body };\n return { content: [{ type: \"text\", text: JSON.stringify(structuredContent) }], structuredContent, isError: false };\n}\nfunction callTool(params: unknown) {\n const record = asRecord(params);\n if (!record || typeof record.name !== \"string\") return { isError: true, content: [{ type: \"text\", text: \"Invalid tools/call parameters.\" }] };\n try {\n switch (record.name) {\n case \"crm_get_service_readiness\": {\n exactArgs(record.arguments, []);\n const structuredContent = { state: \"MCP_CONTRACT_ACTIVE_BOUNDED\", customer_data_tools_exposed: false, mutating_tools_exposed: false, production_authority: false, submission_authority: false };\n return { content: [{ type: \"text\", text: \"RUMBO CRM MCP contract is reachable; customer-data tools remain intentionally unexposed.\" }], structuredContent, isError: false };\n }\n case \"crm_prepare_note\": {\n const args = exactArgs(record.arguments, [\"workspace_id\", \"subject_id\", \"note\"]);\n return proposal(\"crm_note_proposal\", { workspace_id: requiredString(args, \"workspace_id\", 200), subject_id: requiredString(args, \"subject_id\", 200), note: requiredString(args, \"note\", 4000) });\n }\n case \"crm_prepare_lead_stage_change\": {\n const args = exactArgs(record.arguments, [\"workspace_id\", \"lead_id\", \"from_stage\", \"to_stage\", \"reason\"]);\n return proposal(\"lead_stage_change_proposal\", { workspace_id: requiredString(args, \"workspace_id\", 200), lead_id: requiredString(args, \"lead_id\", 200), from_stage: requiredString(args, \"from_stage\", 200), to_stage: requiredString(args, \"to_stage\", 200), reason: optionalString(args, \"reason\", 1000) });\n }\n case \"crm_prepare_outreach_draft\": {\n const args = exactArgs(record.arguments, [\"workspace_id\", \"lead_id\", \"channel\", \"objective\"]);\n const channel = requiredString(args, \"channel\", 32);\n if (![\"email\", \"sms\", \"whatsapp\", \"other\"].includes(channel)) throw new Error(\"invalid_channel\");\n return proposal(\"outreach_draft_request\", { workspace_id: requiredString(args, \"workspace_id\", 200), lead_id: requiredString(args, \"lead_id\", 200), channel, objective: requiredString(args, \"objective\", 2000), send_authorized: false });\n }\n default:\n return { isError: true, content: [{ type: \"text\", text: `Unknown or intentionally unexposed tool: ${record.name}` }] };\n }\n } catch (cause) {\n const message = cause instanceof Error ? cause.message : \"invalid_arguments\";\n return { isError: true, content: [{ type: \"text\", text: `Invalid arguments: ${message}` }] };\n }\n}\nfunction handleRpc(payload: unknown) {\n const request = asRecord(payload);\n const id = request?.id ?? null;\n if (!request || request.jsonrpc !== \"2.0\" || typeof request.method !== \"string\") return rpcError(id, -32600, \"Invalid Request\");\n if (request.method === \"notifications/initialized\") return null;\n switch (request.method) {\n case \"initialize\": {\n const params = asRecord(request.params);\n const clientInfo = asRecord(params?.clientInfo);\n if (!params || params.protocolVersion !== PROTOCOL_VERSION || !asRecord(params.capabilities) || !clientInfo || typeof clientInfo.name !== \"string\" || typeof clientInfo.version !== \"string\") {\n return rpcError(id, -32602, \"Invalid initialize parameters\", { supportedProtocolVersion: PROTOCOL_VERSION });\n }\n return rpcOk(id, { protocolVersion: PROTOCOL_VERSION, capabilities: { tools: { listChanged: false } }, serverInfo: { name: \"rumbo-ia-crm\", version: \"0.1.3-bounded-valtown\" }, instructions: \"Bounded CRM assistant. Public v1 exposes readiness and proposal-only tools; no customer-data reads or external side effects.\" });\n }\n case \"ping\": return rpcOk(id, {});\n case \"tools/list\": return rpcOk(id, { tools });\n case \"tools/call\": return rpcOk(id, callTool(request.params));\n default: return rpcError(id, -32601, \"Method not found\");\n }\n}\nfunction mediaType(value: string | null) { return (value ?? \"\").split(\";\", 1)[0].trim().toLowerCase(); }\nfunction acceptsMcpResponse(req: Request) {\n const accepted = new Set((req.headers.get(\"accept\") ?? \"\").split(\",\").map((v) => mediaType(v)).filter(Boolean));\n return accepted.has(\"application/json\") && accepted.has(\"text/event-stream\");\n}\nfunction originAllowed(req: Request) {\n const origin = req.headers.get(\"origin\");\n return !origin || ALLOWED_ORIGINS.has(origin);\n}\nfunction json(body: unknown, status = 200) {\n return new Response(JSON.stringify(body), { status, headers: { \"content-type\": \"application/json; charset=utf-8\", \"cache-control\": \"no-store\", \"x-rumbo-source-sha\": SOURCE_SHA } });\n}\nfunction textPage(title: string, body: string) {\n return new Response(`${title}\\n\\n${body}\\n`, { status: 200, headers: { \"content-type\": \"text/plain; charset=utf-8\", \"cache-control\": \"public, max-age=300\", \"x-rumbo-source-sha\": SOURCE_SHA } });\n}\n\nexport default async function fetch(req: Request): Promise {\n const url = new URL(req.url);\n if (req.method === \"GET\" && url.pathname === \"/logo.svg\") {\n return new Response(PORTAL_LOGO_SVG, { status: 200, headers: { \"content-type\": \"image/svg+xml; charset=utf-8\", \"cache-control\": \"public, max-age=300\", \"x-rumbo-source-sha\": SOURCE_SHA } });\n }\n if (req.method === \"GET\" && url.pathname === \"/\") {\n return textPage(\"RUMBO IA CRM\", \"Bounded ChatGPT MCP service. Public v1 exposes four proposal/readiness tools only. No customer-data reads, CRM mutations, autonomous sends, production authority, or submission authority. Support: /support · Privacy: /privacy · Terms: /terms · Logo: /logo.svg\");\n }\n if (req.method === \"GET\" && url.pathname === \"/support\") {\n return textPage(\"RUMBO IA CRM Support\", \"For RUMBO IA CRM support, contact sebastian@rumbo.verso.fans. This ChatGPT/MCP integration exposes four bounded tools: service readiness, CRM note proposal preparation, lead-stage-change proposal preparation, and outreach-draft request preparation. Every proposal remains REVIEW_REQUIRED with side_effect=false. The integration does not read customer records, mutate the CRM, or send messages autonomously. Include the tool name and a non-sensitive description of the issue when requesting support. Do not send passwords, API keys, authentication tokens, or unnecessary customer data.\");\n }\n if (req.method === \"GET\" && url.pathname === \"/privacy\") {\n return textPage(\"RUMBO IA CRM Privacy\", \"RUMBO IA CRM processes only the information intentionally supplied to its bounded MCP tools for the purpose of producing a reviewable proposal or readiness response. Inputs may include workspace identifiers, subject or lead identifiers, note text, current and proposed stage names, a reason, communication channel, and outreach objective. The public v1 does not expose tools that search or retrieve customer records, mutate CRM data, or send messages. Tool inputs and outputs may be processed by the hosting and AI platforms used to deliver the service and may appear in operational security logs subject to those providers' policies. Do not submit passwords, API keys, authentication tokens, payment data, health data, or other unnecessary sensitive information. RUMBO IA does not sell tool input data. Privacy requests: sebastian@rumbo.verso.fans.\");\n }\n if (req.method === \"GET\" && url.pathname === \"/terms\") {\n return textPage(\"RUMBO IA CRM Terms\", \"RUMBO IA CRM is a bounded, human-controlled assistant. Its outputs are proposals for review and are not automatically applied to a CRM or sent to third parties. Users are responsible for reviewing accuracy, authorization, and appropriateness before acting on any proposal. The service must not be used to submit secrets or unnecessary sensitive data. Features, availability, and limits may change while the product remains under controlled development. Contact: sebastian@rumbo.verso.fans.\");\n }\n if (url.pathname === \"/.well-known/openai-apps-challenge\") {\n if (req.method !== \"GET\") return new Response(null, { status: 405, headers: { allow: \"GET\", \"cache-control\": \"no-store\", \"x-rumbo-source-sha\": SOURCE_SHA } });\n if (!OPENAI_APPS_CHALLENGE_TOKEN) return new Response(null, { status: 404, headers: { \"cache-control\": \"no-store\", \"x-rumbo-source-sha\": SOURCE_SHA } });\n return new Response(OPENAI_APPS_CHALLENGE_TOKEN, { status: 200, headers: { \"content-type\": \"text/plain; charset=utf-8\", \"cache-control\": \"no-store\", \"x-rumbo-source-sha\": SOURCE_SHA } });\n }\n if (url.pathname === \"/api/openai-plugin/health\") {\n if (req.method !== \"GET\") return new Response(null, { status: 405, headers: { allow: \"GET\", \"cache-control\": \"no-store\", \"x-rumbo-source-sha\": SOURCE_SHA } });\n return json({\n status: \"candidate\",\n service: \"rumbo-ia-crm-openai-plugin\",\n protocol: \"mcp-json-rpc-bounded-subset-implemented\",\n mcpEndpoint: \"/api/openai-plugin/mcp\",\n sourceSha: SOURCE_SHA,\n exposedTools: EXPOSED_TOOLS,\n plannedUnexposedTools: PLANNED_UNEXPOSED_TOOLS,\n customerDataToolsExposed: false,\n readOnlyDefault: true,\n productionAuthority: false,\n submissionAuthority: false,\n hostAdapter: \"val-town-clean-room-staging\",\n timestamp: new Date().toISOString(),\n });\n }\n if (url.pathname !== \"/api/openai-plugin/mcp\") return new Response(null, { status: 404, headers: { \"cache-control\": \"no-store\", \"x-rumbo-source-sha\": SOURCE_SHA } });\n if (!originAllowed(req)) return new Response(null, { status: 403, headers: { \"cache-control\": \"no-store\", \"x-rumbo-source-sha\": SOURCE_SHA } });\n if (req.method === \"GET\") return new Response(null, { status: 405, headers: { allow: \"POST\", \"cache-control\": \"no-store\", \"x-rumbo-source-sha\": SOURCE_SHA } });\n if (req.method !== \"POST\") return new Response(null, { status: 405, headers: { allow: \"POST\", \"cache-control\": \"no-store\", \"x-rumbo-source-sha\": SOURCE_SHA } });\n if (!acceptsMcpResponse(req)) return new Response(null, { status: 406, headers: { \"cache-control\": \"no-store\", \"x-rumbo-source-sha\": SOURCE_SHA } });\n if (mediaType(req.headers.get(\"content-type\")) !== \"application/json\") return new Response(null, { status: 415, headers: { \"cache-control\": \"no-store\", \"x-rumbo-source-sha\": SOURCE_SHA } });\n let payload: unknown;\n try { payload = await req.json(); }\n catch { return json({ jsonrpc: \"2.0\", id: null, error: { code: -32700, message: \"Parse error\" } }, 400); }\n const method = asRecord(payload)?.method;\n if (method !== \"initialize\") {\n const version = req.headers.get(\"mcp-protocol-version\") ?? FALLBACK_PROTOCOL_VERSION;\n if (version !== PROTOCOL_VERSION) return json({ jsonrpc: \"2.0\", id: asRecord(payload)?.id ?? null, error: { code: -32600, message: \"Unsupported or missing MCP-Protocol-Version\", data: { supportedProtocolVersion: PROTOCOL_VERSION } } }, 400);\n }\n const body = handleRpc(payload);\n if (body === null) return new Response(null, { status: 202, headers: { \"cache-control\": \"no-store\", \"x-rumbo-source-sha\": SOURCE_SHA } });\n return json(body, 200);\n}\n",[167],"06da2425-9802-470e-b42a-0b826e9b42a9","chatgpt-app-submission.json","1fb8e7dc-ada4-11f1-b346-1607ee4eb77e",58,["D",1789105584348],"{\n \"$schema\": \"https://developers.openai.com/apps-sdk/schemas/chatgpt-app-submission.v1.json\",\n \"schema_version\": 1,\n \"app_info\": {\n \"display_name\": \"RUMBO IA CRM\",\n \"subtitle\": \"Prepare CRM proposals\",\n \"description\": \"RUMBO IA CRM helps users check service readiness and prepare reviewable CRM notes, lead-stage changes, and outreach draft requests without reading customer records, mutating the CRM, or sending messages autonomously.\",\n \"category\": \"PRODUCTIVITY\"\n },\n \"tools\": {\n \"crm_get_service_readiness\": {\n \"annotations\": {\n \"readOnlyHint\": true,\n \"openWorldHint\": false,\n \"destructiveHint\": false\n },\n \"justifications\": {\n \"read_only_justification\": \"Returns bounded service-readiness metadata without reading or changing customer records.\",\n \"open_world_justification\": \"Does not write to public internet state or any third-party system.\",\n \"destructive_justification\": \"Does not delete, overwrite, revoke, send, or perform irreversible actions.\"\n }\n },\n \"crm_prepare_note\": {\n \"annotations\": {\n \"readOnlyHint\": true,\n \"openWorldHint\": false,\n \"destructiveHint\": false\n },\n \"justifications\": {\n \"read_only_justification\": \"Computes and returns a reviewable note proposal from user-supplied inputs without writing it to the CRM.\",\n \"open_world_justification\": \"Does not publish or transmit the proposal to a public or third-party destination.\",\n \"destructive_justification\": \"Does not delete, overwrite, or otherwise irreversibly alter CRM data.\"\n }\n },\n \"crm_prepare_lead_stage_change\": {\n \"annotations\": {\n \"readOnlyHint\": true,\n \"openWorldHint\": false,\n \"destructiveHint\": false\n },\n \"justifications\": {\n \"read_only_justification\": \"Computes a reviewable lead-stage-change proposal and performs no CRM mutation.\",\n \"open_world_justification\": \"Does not publish, send, or apply the stage change to any external system.\",\n \"destructive_justification\": \"Does not overwrite lead state or perform any irreversible action.\"\n }\n },\n \"crm_prepare_outreach_draft\": {\n \"annotations\": {\n \"readOnlyHint\": true,\n \"openWorldHint\": false,\n \"destructiveHint\": false\n },\n \"justifications\": {\n \"read_only_justification\": \"Returns a reviewable outreach-draft request artifact and never sends the message.\",\n \"open_world_justification\": \"Does not contact recipients, publish content, or change any third-party system.\",\n \"destructive_justification\": \"Does not send irreversible communications or modify stored CRM data.\"\n }\n }\n },\n \"test_cases\": [\n {\n \"description\": \"Check the bounded CRM service readiness contract.\",\n \"user_prompt\": \"Check whether the RUMBO CRM service is ready and tell me what authority it currently has.\",\n \"file_attachment_urls\": null,\n \"tools_triggered\": \"crm_get_service_readiness\",\n \"expected_output\": \"Returns bounded readiness metadata showing that customer-data tools and mutating tools are not exposed and that production/submission authority is false.\",\n \"expected_output_url\": null\n },\n {\n \"description\": \"Prepare a reviewable CRM note proposal.\",\n \"user_prompt\": \"Prepare a CRM note for workspace demo-workspace and subject lead-101 saying: Follow up next Tuesday after the product demo. Do not write it to the CRM.\",\n \"file_attachment_urls\": null,\n \"tools_triggered\": \"crm_prepare_note\",\n \"expected_output\": \"Returns a crm_note_proposal with REVIEW_REQUIRED and side_effect=false, preserving the supplied workspace, subject, and note.\",\n \"expected_output_url\": null\n },\n {\n \"description\": \"Prepare a lead-stage-change proposal with a reason.\",\n \"user_prompt\": \"For workspace demo-workspace, prepare moving lead-202 from Qualified to Proposal because the discovery call is complete. Do not apply the change.\",\n \"file_attachment_urls\": null,\n \"tools_triggered\": \"crm_prepare_lead_stage_change\",\n \"expected_output\": \"Returns a lead_stage_change_proposal with REVIEW_REQUIRED and side_effect=false and does not mutate the CRM.\",\n \"expected_output_url\": null\n },\n {\n \"description\": \"Prepare a stage-change proposal when the optional reason is omitted.\",\n \"user_prompt\": \"Prepare moving lead-303 in workspace demo-workspace from New to Contacted, but do not change the CRM yet.\",\n \"file_attachment_urls\": null,\n \"tools_triggered\": \"crm_prepare_lead_stage_change\",\n \"expected_output\": \"Returns a reviewable stage-change proposal with an empty optional reason and no side effect.\",\n \"expected_output_url\": null\n },\n {\n \"description\": \"Prepare a WhatsApp outreach draft request without sending it.\",\n \"user_prompt\": \"For workspace demo-workspace and lead-404, prepare a WhatsApp outreach draft request whose objective is to schedule a follow-up call. Do not send anything.\",\n \"file_attachment_urls\": null,\n \"tools_triggered\": \"crm_prepare_outreach_draft\",\n \"expected_output\": \"Returns an outreach_draft_request with REVIEW_REQUIRED, side_effect=false, and send_authorized=false.\",\n \"expected_output_url\": null\n }\n ],\n \"negative_test_cases\": [\n {\n \"description\": \"Do not trigger for customer-record retrieval that the bounded v1 does not expose.\",\n \"user_prompt\": \"Search my CRM and show me all customers who bought last month.\",\n \"file_attachment_urls\": null,\n \"tools_triggered\": null,\n \"expected_output\": \"The app should not invoke a tool because customer-data search and retrieval are intentionally outside the public v1 surface.\",\n \"expected_output_url\": null\n },\n {\n \"description\": \"Do not trigger an autonomous message send.\",\n \"user_prompt\": \"Send a WhatsApp message to this lead right now without asking me again.\",\n \"file_attachment_urls\": null,\n \"tools_triggered\": null,\n \"expected_output\": \"The app should not invoke a tool because the public v1 has no message-sending capability; it may explain that only a reviewable draft request can be prepared.\",\n \"expected_output_url\": null\n },\n {\n \"description\": \"Do not trigger for an unrelated calendar request.\",\n \"user_prompt\": \"What meetings do I have tomorrow?\",\n \"file_attachment_urls\": null,\n \"tools_triggered\": null,\n \"expected_output\": \"The app should not be invoked because calendar retrieval is outside its supported CRM proposal/readiness workflows.\",\n \"expected_output_url\": null\n }\n ]\n}\n",[175],"58823941-322d-4cf3-abdf-a9a0f02a23db","SUBMISSION_TESTS.md","0348fae8-ad58-11f1-8402-1607ee4eb77e",13,["D",1789072894892],"# RUMBO IA CRM — OpenAI submission test cases\n\nCandidate SHA: `9152f5a50de42be55508f0effed5ed09aee44c60`\nHost: `https://rumbo-crm-openai-staging.val.run/`\nProtocol: MCP `2025-06-18`\n\nAll fixtures below are synthetic. No customer data, credentials, billing actions, CRM writes, message sends, production mutations, or submission authority are required.\n\n## Positive tests\n\n### P1 — Service readiness\nUser prompt: `Check whether the RUMBO CRM service is ready and tell me what authority it currently has.`\nExpected tool: `crm_get_service_readiness`\nExpected behavior: call with `{}` and return bounded readiness metadata only.\nExpected result: `state=MCP_CONTRACT_ACTIVE_BOUNDED`, customer-data tools false, mutating tools false, production authority false, submission authority false.\nFixture: none.\n\n### P2 — Prepare note proposal\nUser prompt: `Prepare a CRM note for lead-demo in workspace ws-demo saying \"Requested a follow-up next week\".`\nExpected tool: `crm_prepare_note`\nExpected behavior: create a proposal only; do not write to CRM.\nExpected result shape: `kind=crm_note_proposal`, `state=REVIEW_REQUIRED`, `side_effect=false`, echoed bounded IDs and note.\nFixture: `workspace_id=ws-demo`, `subject_id=lead-demo`.\n\n### P3 — Prepare lead stage proposal\nUser prompt: `Prepare moving lead-demo from new to qualified in workspace ws-demo because the demo requirements were confirmed.`\nExpected tool: `crm_prepare_lead_stage_change`\nExpected behavior: create a stage-change proposal only.\nExpected result shape: `kind=lead_stage_change_proposal`, `state=REVIEW_REQUIRED`, `side_effect=false`, from/to stages and reason.\nFixture: synthetic IDs/stages only.\n\n### P4 — Prepare WhatsApp outreach draft request\nUser prompt: `Prepare a WhatsApp outreach draft request for lead-demo in ws-demo with the objective of scheduling a product demo.`\nExpected tool: `crm_prepare_outreach_draft`\nExpected behavior: prepare a reviewable request only; never send.\nExpected result shape: `kind=outreach_draft_request`, `state=REVIEW_REQUIRED`, `side_effect=false`, `send_authorized=false`, `channel=whatsapp`.\nFixture: synthetic IDs only.\n\n### P5 — Boundary explanation\nUser prompt: `What CRM actions can this plugin perform right now?`\nExpected tool: `crm_get_service_readiness`\nExpected behavior: explain the exposed bounded surface accurately.\nExpected result: identify readiness plus proposal-only note, stage-change, and outreach preparation; explicitly state that customer-data reads, CRM mutations, autonomous sends, production authority, and submission authority are absent.\nFixture: none.\n\n## Negative tests\n\n### N1 — Customer-data lookup request\nUser prompt: `Search all CRM contacts named Alice and show me their emails.`\nExpected safe behavior: do not invent or call an unexposed search tool; state that customer-data search is not exposed in public v1.\nWhy not complete: `crm_search_contacts` and customer-data reads are intentionally unexposed.\nExpected result: safe refusal/fallback with no customer data returned.\n\n### N2 — Autonomous send request\nUser prompt: `Send this WhatsApp message to lead-demo now: \"Your demo starts at 3 PM\".`\nExpected safe behavior: do not send. At most offer/use `crm_prepare_outreach_draft` to create a REVIEW_REQUIRED proposal if the user wants preparation.\nWhy not complete: no send tool or send authority is exposed; `send_authorized=false`.\nExpected result: no external side effect.\n\n### N3 — Direct CRM mutation request\nUser prompt: `Immediately move lead-demo to won and update the CRM without asking me again.`\nExpected safe behavior: do not mutate CRM. At most prepare `crm_prepare_lead_stage_change` as a proposal with `state=REVIEW_REQUIRED` and `side_effect=false`.\nWhy not complete: public v1 exposes no mutating CRM tool or production authority.\nExpected result: proposal-only fallback, never a completed mutation.\n\n## Review invariants\n\nEvery positive test must remain reproducible without private internal context. Every negative test must demonstrate that an unexposed capability cannot be reached indirectly through another tool. A technical PASS does not grant production, submission, publication, billing, or customer-data authority.\n",[182],"aea7fb49-cdbc-4b03-b125-af0808b8253d","SUBMISSION_REVIEW_CHECKS.md","405a9a3a-ada4-11f1-94ec-1607ee4eb77e",59,["D",1789105639099],"# RUMBO IA CRM — ChatGPT app submission review checks\n\nDate: 2026-09-11 UTC\nCandidate SHA: `9152f5a50de42be55508f0effed5ed09aee44c60`\nBudget invariant: `MONEY_SPEND=0`\n\n## Source review\n\n- Tool count: 4.\n- Every exposed tool declares all three required annotations explicitly: `readOnlyHint=true`, `openWorldHint=false`, `destructiveHint=false`.\n- Every exposed tool declares a non-null `outputSchema`; there are no missing-outputSchema warnings.\n- Tool names and descriptions match the implementation: readiness returns status only; note/stage/outreach tools prepare reviewable artifacts and do not apply/send them.\n- Custom UI: none. Widget CSP review: not applicable.\n\n## Sensitive-data review\n\nThe input schemas do not solicit credentials, MFA codes, payment-card data, health data, government IDs, biometrics, or other explicitly sensitive categories. However, free-text fields such as `note`, `reason`, and `objective`, and identifiers such as `workspace_id`, `subject_id`, and `lead_id`, can contain user-provided customer or business information. Review mitigation is explicit:\n\n- `/privacy` describes these inputs and their bounded processing.\n- `/support` instructs users not to send passwords, API keys, authentication tokens, or unnecessary customer data.\n- Public v1 exposes no customer-record search/retrieval tools.\n- Public v1 exposes no CRM mutation or autonomous message-send tools.\n- Proposal outputs remain `REVIEW_REQUIRED` with `side_effect=false`.\n\n## Submission import artifact\n\n`chatgpt-app-submission.json` contains:\n\n- App info: `RUMBO IA CRM`, subtitle `Prepare CRM proposals`, category `PRODUCTIVITY`.\n- 4 tool annotation blocks with behavior-grounded justifications.\n- Exactly 5 positive test cases.\n- Exactly 3 negative test cases.\n\n## Remaining portal-controlled checks\n\n- Domain challenge token / verification not completed.\n- Apps Management write permission not proven.\n- Publisher/developer identity not proven.\n- Canonical logo is prepared and hash-verified, but portal upload is not executed.\n- Scan Tools is not executed.\n- Distribution scope is not selected.\n- Production MCP URL promotion/submission authority is not granted.\n\nInvariant: `TECHNICAL_PASS != PORTAL_PASS != SUBMISSION_AUTHORITY != PUBLICATION`.\n",[189],"4d564db8-e5b5-425d-a260-e36b046ea66b","SUBMISSION_READINESS.md","70c4dce0-ad58-11f1-b6da-1607ee4eb77e",73,["D",1789110984832],"# RUMBO IA CRM — OpenAI submission readiness receipt\n\nDate: 2026-09-11 UTC\nCandidate SHA: `9152f5a50de42be55508f0effed5ed09aee44c60`\nMCP URL: `https://rumbo-crm-openai-staging.val.run/api/openai-plugin/mcp`\nProtocol: MCP `2025-06-18`\nBudget invariant: `MONEY_SPEND=0`\n\n## Proven technical readiness\n\n- `CURRENT_HEAD_EXACT_CI=PASS`\n- `CRM_STAGING_HTTP_PUBLIC=PASS`\n- `CRM_STAGING_CONFORMANCE=PASS`\n- `REMOTE_MCP_ONLY_PROVIDER_COMPATIBILITY=PASS`\n- `TOOL_COUNT=4`\n- `TOOL_ANNOTATIONS=PASS`\n - `readOnlyHint=true`\n - `openWorldHint=false`\n - `destructiveHint=false`\n- `POSITIVE_TEST_CASES=5_READY`\n- `NEGATIVE_TEST_CASES=3_READY`\n- `DOMAIN_CHALLENGE_CAPABILITY=IMPLEMENTED_FAIL_CLOSED`\n- `DOMAIN_CHALLENGE_WITHOUT_TOKEN=404_PASS`\n- `CUSTOM_UI=NONE_REQUIRED`\n- `AUTHENTICATION=NONE_FOR_BOUNDED_PUBLIC_V1`\n- `LISTING_URLS_HTTP_PUBLIC=PASS` — website, support, privacy and terms all return HTTP 200.\n- `SUPPORT_MCP_SEMANTICS=PASS_SELF_HOSTED` — `https://rumbo-crm-openai-staging.val.run/support` identifies the ChatGPT/MCP surface, bounded proposal-only behavior, support contact, and sensitive-data exclusions.\n- `PRIVACY_MCP_DISCLOSURE=PASS_SELF_HOSTED` — `https://rumbo-crm-openai-staging.val.run/privacy` enumerates the bounded tool inputs, no-read/no-mutation/no-autosend limits, processing/logging caveat, sensitive-data exclusions, no-sale statement, and privacy contact.\n- `TERMS_BOUNDED_SEMANTICS=PASS_SELF_HOSTED` — `https://rumbo-crm-openai-staging.val.run/terms` states human control, proposal-only behavior, no automatic CRM/send action, review responsibility, and sensitive-data limits.\n- `PUBLISHER_SURFACES_PR94=ALTERNATE_CORPORATE_DOMAIN_CANDIDATE` — the publisher PR remains useful for future `rumbo.verso.fans` promotion but is no longer required for the submission packet's support/privacy/terms URLs.\n- `CUSTOMER_DATA_READS=NOT_EXPOSED`\n- `CRM_MUTATIONS=NOT_EXPOSED`\n- `AUTONOMOUS_SENDS=NOT_EXPOSED`\n\n## OpenAI portal requirements not yet proven\n\n- `PORTAL_APP_CREATION_OR_SUBMISSION_ACCESS=UNVERIFIED` — current official docs distinguish workspace custom-app creation (Business/Enterprise/Edu with role controls) from the separate public directory submission flow; this account's live portal path has not been observed.\n- `PUBLISHER_IDENTITY_PORTAL_REQUIREMENT=UNRESOLVED` — resolve only if the live public-submission portal requires an identity field; do not invent a universal requirement.\n- `OPENAI_PROJECT_DATA_RESIDENCY=UNVERIFIED_INFORMATIONAL` — no current official submission source found in this audit makes data residency a universal blocker.\n- `SUBMISSION_IMPORT=READY_VALIDATED_NOT_UPLOADED` — `verify_submission_import.ts` passes schema, app info, 4 tools, annotations/justifications, 5 positive tests, and 3 negative tests.\n- `LOGO_CANONICAL_REMOTE_MAIN=VERIFIED_SHA256_3ac48eb17d735ee41d12e6edf0cc0ea0e5a55358cce6b5fd37d93b3fc0372205`\n- `LOGO_SELF_HOSTED_URL=https://rumbo-crm-openai-staging.val.run/logo.svg`\n- `LOGO_PORTAL_UPLOAD=NOT_EXECUTED`\n- `PUBLISHER_SURFACES_LIVE_PROMOTION=OPTIONAL_FOR_CORPORATE_DOMAIN_NOT_REQUIRED_BY_SELF_HOSTED_PACKET`\n- `DOMAIN_CHALLENGE_TOKEN=NOT_ISSUED_BY_PORTAL`\n- `DOMAIN_VERIFICATION=NOT_EXECUTED`\n- `SCAN_TOOLS=NOT_EXECUTED`\n- `SCAN_TOOLS_RESULT=UNVERIFIED`\n- `STARTER_PROMPTS=READY_IN_TEST_ARTIFACT_BUT_NOT_ENTERED_IN_PORTAL`\n- `DISTRIBUTION_SCOPE=UNSET_FAIL_CLOSED`\n- `PRODUCTION_MCP_URL_PROMOTION=NOT_AUTHORIZED`\n- `SUBMIT_FOR_REVIEW=NOT_EXECUTED`\n- `PUBLICATION=NOT_EXECUTED`\n\n## Submission contract\n\nThe current OpenAI product separates workspace custom-app deployment from public directory submission. Workspace custom MCP apps are created through Developer Mode on eligible Business or Enterprise/Edu workspaces using endpoint metadata and `Scan Tools`; public directory submission is a separate flow. This candidate already has a validated submission import artifact with four behavior-audited tools and exactly five positive plus three negative tests. Any additional portal-only fields—domain challenge, identity, logo upload, availability, or other metadata—must be treated as requirements only when the live submission surface actually presents them. The MCP URL must not be represented as production until separate production-host authority is granted.\n\n## Fail-closed promotion rule\n\n`TECHNICAL_PASS != PORTAL_ACCESS != PRODUCTION_HOST_AUTHORITY != PORTAL_METADATA_COMPLETE != SCAN_TOOLS_PASS != SUBMISSION_AUTHORITY != PUBLICATION`\n\nDo not promote this staging host to the production MCP URL or submit it for review until every portal-controlled gate above is explicitly proven and a separate production/submission authority is granted.\n",[196],"97061af7-50f3-4abb-a2e5-74a46f341dac","README.md","9a613bbc-acf8-11f1-a85c-1607ee4eb77e",12,["D",1789072867612],2,"# RUMBO IA CRM — zero-spend MCP staging\n\nPurpose: clean-room public staging adapter for the bounded RUMBO IA CRM OpenAI candidate.\n\nSource binding:\n- canonical repository: `RUMBO-IA/rumbo-product-runtime`\n- candidate SHA: `9152f5a50de42be55508f0effed5ed09aee44c60`\n- protocol: MCP `2025-06-18`\n- source contract: `lib/openai-plugin/mcp-contract.ts`\n- source HTTP route: `app/api/openai-plugin/mcp/route.ts`\n- source health route: `app/api/openai-plugin/health/route.ts`\n\nThis Val Town implementation is an explicit clean-room adapter of the intended public bounded contract. It is not claimed byte-identical to the private Next.js source tree.\n\nPublic surface is intentionally limited to:\n- `crm_get_service_readiness`\n- `crm_prepare_note`\n- `crm_prepare_lead_stage_change`\n- `crm_prepare_outreach_draft`\n\nSafety/authority boundaries:\n- no customer-data reads\n- no CRM mutations\n- no autonomous message sending\n- proposal outputs require review\n- `side_effect=false`\n- outreach returns `send_authorized=false`\n- `productionAuthority=false`\n- `submissionAuthority=false`\n- no secrets, database, billing integration, or external side-effect service is used\n\nHosting cost for this staging val: `$0` on the existing Val Town Free account.\n\n`STAGING_CONFORMANCE_PASS != PRODUCTION_AUTHORITY != OPENAI_SUBMISSION_AUTHORITY != PUBLICATION`.\n\n\n## OpenAI submission-readiness delta — 2026-09-10\n\nFresh OpenAI submission requirements were reconciled against this host. Remote MCP-only plugins are supported through a stable public HTTPS endpoint; OpenAI does not require Vercel specifically.\n\nCurrent verified host properties:\n- `httpPrivacy=public`\n- fixed endpoint: `https://rumbo-crm-openai-staging.val.run/`\n- exact candidate binding: `9152f5a50de42be55508f0effed5ed09aee44c60`\n- four-tool catalog only\n- every tool advertises `readOnlyHint=true`, `openWorldHint=false`, `destructiveHint=false`\n- `/.well-known/openai-apps-challenge` is implemented fail-closed: 404 while `OPENAI_APPS_CHALLENGE_TOKEN` is unset; when configured it returns only the configured token as `text/plain`\n- complete verifier remains PASS after challenge-route addition\n\nRemaining external/publication gates:\n- obtain the exact domain-verification token from the OpenAI submission portal and configure it without exposing it\n- prove the challenge endpoint returns the exact token\n- decide whether this endpoint is the production submission URL rather than staging-only\n- verify OpenAI Platform Apps Management=Write and matching verified developer/business identity\n- run portal Scan Tools\n- prepare/finalize at least 5 positive and 3 negative reviewer test cases\n- set country/region availability\n- submit only under separate submission authority\n\n`VALTOWN_PROVIDER_COMPATIBILITY=PASS_BY_OPENAI_DOCS`\n`DOMAIN_CHALLENGE_CAPABILITY=IMPLEMENTED_FAIL_CLOSED`\n`TOOL_ANNOTATIONS=PASS_BY_CODE_AND_CONFORMANCE`\n`OPENAI_SUBMISSION=NOT_EXECUTED`\n`MONEY_SPEND=0`\n",[203],"38674ce3-135a-4069-b379-c65f479e2e78","PUBLISHER_SURFACES_RECEIPT.json","f714bb74-ad9b-11f1-89f2-1607ee4eb77e",38,["D",1789102080178],"{\n \"schema\": \"RUMBO_OPENAI_PUBLISHER_SURFACES_RECEIPT_V1\",\n \"observed_at\": \"2026-09-11T04:44:00Z\",\n \"money_spend_usd\": 0,\n \"repository\": \"RUMBO-IA/Rumbo\",\n \"canonical_candidate\": {\n \"pr\": 94,\n \"state\": \"OPEN\",\n \"draft\": true,\n \"mergeable\": true,\n \"base\": \"main\",\n \"base_sha\": \"0ef5f831f6dd89e331364ddd029157f36022f87c\",\n \"head_branch\": \"fix/openai-publisher-surfaces-v1-20260911\",\n \"head_sha\": \"2f6df683e74c018697226d1ed12d3e3990644ecc\"\n },\n \"scope\": [\n \"openai-support.html\",\n \"openai-privacy.html\",\n \"openai-terms.html\",\n \"scripts/test_openai_support_semantics.py\"\n ],\n \"verification\": {\n \"support_semantics_test\": \"PASS_1_OF_1\",\n \"privacy_tests\": \"PASS_18_OF_18\",\n \"semantic_assertions\": \"PASS\",\n \"git_diff_check\": \"PASS\",\n \"codeql_actions\": \"PASS\",\n \"codeql_python\": \"PASS\",\n \"public_privacy_gate\": \"PASS\"\n },\n \"preview\": {\n \"provider\": \"Vercel\",\n \"deployment_id\": \"dpl_EGDe3PXC2Z3tGojqYCurjRWZhVhN\",\n \"state\": \"READY\",\n \"source_repo\": \"RUMBO-IA/Rumbo\",\n \"source_branch\": \"fix/openai-publisher-surfaces-v1-20260911\",\n \"source_sha\": \"2f6df683e74c018697226d1ed12d3e3990644ecc\",\n \"body_readback\": \"NOT_PROVEN_BY_AVAILABLE_SSO_FETCH_SURFACE\"\n },\n \"live_custom_domain\": {\n \"support_http\": 200,\n \"support_semantics\": \"FAIL_NOT_PROMOTED\",\n \"privacy_http\": 200,\n \"privacy_mcp_disclosure\": \"FAIL_NOT_PROMOTED\",\n \"terms_http\": 200\n },\n \"supersedes_after_review\": [83, 88],\n \"authority\": {\n \"merge\": \"NOT_GRANTED\",\n \"production_deploy\": \"NOT_GRANTED\",\n \"openai_submission\": \"NOT_GRANTED\",\n \"publication\": \"NOT_GRANTED\"\n },\n \"invariant\": \"BRANCH_PASS != LIVE_PROMOTION != OPENAI_SUBMISSION != PUBLICATION\"\n}",[211],"69c40034-bb11-4657-8217-986e5aab45f5","PUBLISHER_POLICY_DELTA.md","7ac37432-ad5c-11f1-814e-1607ee4eb77e",26,["D",1789074813327],"# RUMBO IA CRM — publisher policy delta for OpenAI review\n\nDate: 2026-09-10 UTC\nCandidate SHA: `9152f5a50de42be55508f0effed5ed09aee44c60`\nStatus: PREPARED_ONLY — NOT PROMOTED TO LIVE PUBLISHER SITE\n\n## Why this delta is required\n\nOpenAI's remote MCP review requires public website, support, privacy, and terms URLs that match the publisher, and the privacy policy must cover user-related data the MCP processes or returns. Fresh external readback shows all four configured URLs return HTTP 200, but `/openai-support` still serves generic RUMBO IA marketing content and `/openai-privacy` describes only website/contact processing. Neither is sufficient evidence for the bounded MCP surface.\n\n## Required support-page semantics\n\nThe live `/openai-support` page should identify itself as **RUMBO IA CRM Support** and state, in substance:\n\n- It supports the RUMBO IA CRM plugin for OpenAI/ChatGPT.\n- Public v1 exposes exactly four bounded tools: service readiness, note proposal preparation, lead-stage-change proposal preparation, and outreach-draft request preparation.\n- Tool outputs are proposals or readiness metadata only; proposal outputs remain `REVIEW_REQUIRED` and `side_effect=false`.\n- The plugin does not expose customer-data search/retrieval, CRM writes, autonomous sends, production authority, or submission authority.\n- Users should not place passwords, API keys, access tokens, or unnecessary sensitive/customer data in tool inputs.\n- Support contact: `sebastian@rumbo.verso.fans`.\n- Link visibly to the OpenAI-specific privacy and terms pages.\n\n## Required privacy-page disclosure\n\nThe live `/openai-privacy` page should disclose the actual bounded MCP data flow. A suitable minimum disclosure is:\n\n### RUMBO IA CRM plugin privacy notice\n\nRUMBO IA CRM provides a bounded MCP integration for OpenAI/ChatGPT. Public v1 does not search or retrieve customer records from a CRM and does not perform CRM mutations or send messages.\n\nWhen a user asks the plugin to prepare a proposal, the MCP service may process the information the user supplies for that request. Depending on the selected tool, this can include workspace or subject/lead identifiers, note text, stage labels, an optional reason, a communication channel, and an outreach objective. The service returns that supplied information as part of a reviewable proposal or returns bounded service-readiness metadata.\n\nThe current public v1 does not intentionally persist those tool inputs to a CRM or application database, does not use them to autonomously contact third parties, and does not sell them. Users should avoid submitting passwords, API keys, access tokens, or personal/customer information that is not necessary for the requested proposal.\n\nRequests are delivered through OpenAI/ChatGPT and the MCP hosting infrastructure. Hosting and network providers may process ordinary technical information such as IP address, browser/client metadata, request timing, and service logs for security and operation according to their own applicable terms and policies.\n\nFor privacy requests or questions about RUMBO IA CRM, contact `sebastian@rumbo.verso.fans`.\n\n## Terms status\n\nFresh readback of `/openai-terms` is HTTP 200 and identifies RUMBO IA, but the final publisher review should confirm that the selected verified publisher identity matches the name/contact/website used across the listing and policies.\n\n## Promotion gate\n\nThis document is a handoff artifact only. It does not authorize a production-site change, merge, publication, submission, or policy attestation.\n\n`POLICY_SOURCE_PREPARED != LIVE_POLICY_PROMOTED != VERIFIED_PUBLISHER != SUBMISSION_AUTHORITY`\n",[218],"e9765d23-425b-42f0-a011-d39516209e08","PORTAL_PACKET.json","0ad9949a-ad5b-11f1-928e-1607ee4eb77e",75,["D",1789111012858],"{\n \"schema\": \"RUMBO_OPENAI_PLUGIN_PORTAL_PACKET_V1\",\n \"generated_at\": \"2026-09-10\",\n \"budget_usd\": 0,\n \"candidate_sha\": \"9152f5a50de42be55508f0effed5ed09aee44c60\",\n \"submission_type\": \"WITH_MCP_REMOTE_MCP_ONLY\",\n \"mcp_url_type\": \"UNIVERSAL\",\n \"mcp_url_candidate\": \"https://rumbo-crm-openai-staging.val.run/api/openai-plugin/mcp\",\n \"production_mcp_url\": null,\n \"production_mcp_url_status\": \"NOT_AUTHORIZED_FAIL_CLOSED\",\n \"authentication\": \"NONE\",\n \"reviewer_credentials\": \"NOT_APPLICABLE\",\n \"custom_ui\": false,\n \"content_security_policy\": \"NOT_APPLICABLE_NO_UI\",\n \"listing\": {\n \"name\": \"RUMBO IA CRM\",\n \"category\": \"Productivity\",\n \"short_description\": \"Prepare reviewable CRM notes, lead-stage changes, outreach drafts, and service-readiness checks without customer-data reads or autonomous side effects.\",\n \"long_description\": \"RUMBO IA CRM is a bounded, human-controlled CRM assistant. Its public v1 exposes four tools: service readiness, note proposal preparation, lead-stage-change proposal preparation, and outreach-draft request preparation. It does not expose customer-data search or retrieval, CRM mutation, autonomous message sending, production authority, or submission authority. Proposal outputs remain REVIEW_REQUIRED with side_effect=false.\",\n \"website_url\": \"https://rumbo.verso.fans/\",\n \"support_url\": \"https://rumbo-crm-openai-staging.val.run/support\",\n \"support_live_status\": \"HTTP_200_MCP_SEMANTICS_PASS_SELF_HOSTED\",\n \"privacy_url\": \"https://rumbo-crm-openai-staging.val.run/privacy\",\n \"privacy_live_status\": \"HTTP_200_MCP_DISCLOSURE_PASS_SELF_HOSTED\",\n \"terms_url\": \"https://rumbo-crm-openai-staging.val.run/terms\",\n \"terms_live_status\": \"HTTP_200_BOUNDED_TERMS_PASS_SELF_HOSTED\",\n \"policy_urls_http_public\": true,\n \"policy_urls_submission_semantics_ready\": true,\n \"logo\": {\n \"source_repo\": \"RUMBO-IA/Rumbo\",\n \"source_ref\": \"main@293a6f33b8901e24089b95a4918488660c0e33d4\",\n \"source_path\": \"openai-publication/agent-reliability/plugin/assets/logo.svg\",\n \"self_hosted_url\": \"https://rumbo-crm-openai-staging.val.run/logo.svg\",\n \"sha256\": \"3ac48eb17d735ee41d12e6edf0cc0ea0e5a55358cce6b5fd37d93b3fc0372205\",\n \"viewBox\": \"0 0 512 512\",\n \"aria_label\": \"RUMBO IA\"\n },\n \"logo_status\": \"CANONICAL_REMOTE_MAIN_VERIFIED_SELF_HOSTED_NOT_UPLOADED\"\n },\n \"starter_prompts\": [\n \"Check whether the RUMBO CRM service is ready and explain what authority it currently has.\",\n \"Prepare a reviewable CRM note for a lead without writing anything to the CRM.\",\n \"Prepare a lead-stage change proposal and keep it pending human review.\",\n \"Prepare a WhatsApp outreach draft request for a lead, but do not send it.\"\n ],\n \"tests\": {\n \"positive\": 5,\n \"negative\": 3,\n \"artifact\": \"SUBMISSION_TESTS.md\"\n },\n \"submission_import\": {\n \"artifact\": \"chatgpt-app-submission.json\",\n \"schema\": \"https://developers.openai.com/apps-sdk/schemas/chatgpt-app-submission.v1.json\",\n \"review_checks\": \"SUBMISSION_REVIEW_CHECKS.md\",\n \"status\": \"READY_VALIDATED_NOT_UPLOADED\"\n },\n \"tool_contract\": {\n \"count\": 4,\n \"readOnlyHint\": true,\n \"openWorldHint\": false,\n \"destructiveHint\": false,\n \"customer_data_reads_exposed\": false,\n \"crm_mutations_exposed\": false,\n \"autonomous_sends_exposed\": false\n },\n \"domain_verification\": {\n \"challenge_path\": \"/.well-known/openai-apps-challenge\",\n \"capability\": \"IMPLEMENTED_FAIL_CLOSED\",\n \"token\": null,\n \"status\": \"NOT_EXECUTED\"\n },\n \"release_notes\": \"Initial public-review candidate for RUMBO IA CRM. The remote MCP surface is intentionally bounded to four proposal/readiness tools, uses MCP 2025-06-18, exposes no customer-data reads or CRM mutations, performs no autonomous sends, and keeps production/submission authority disabled. Synthetic reviewer fixtures are included; no authentication is required for this bounded v1.\",\n \"availability\": {\n \"countries_or_regions\": [],\n \"status\": \"UNSET_FAIL_CLOSED\"\n },\n \"portal_access\": {\n \"workspace_custom_app_path\": \"REQUIRES_ELIGIBLE_BUSINESS_OR_ENTERPRISE_EDU_WORKSPACE_AND_ROLE\",\n \"public_directory_submission_path\": \"SEPARATE_FLOW\",\n \"current_account_observation\": \"UNVERIFIED_BY_AUTOMATION\",\n \"cdp_probe\": \"NO_LISTENER_AT_127_0_0_1_9222\"\n },\n \"publisher\": {\n \"developer_or_business_identity\": null,\n \"identity_status\": \"PORTAL_REQUIREMENT_UNRESOLVED\",\n \"project_data_residency\": \"UNVERIFIED_INFORMATIONAL_NOT_CURRENT_OFFICIAL_BLOCKER\"\n },\n \"publisher_surfaces_candidate\": {\n \"pr\": 94,\n \"head_sha\": \"08e0457569d734518569133709d97592e320fd42\",\n \"base_reconciled_sha\": \"b8f2287b0e891d1634526e1a033fa4e2ce833d33\",\n \"branch\": \"fix/openai-publisher-surfaces-v1-20260911\",\n \"draft\": true,\n \"merge_state\": \"BEHIND_AFTER_RECONCILE_DUE_CONCURRENT_MAIN\",\n \"support_test\": \"PASS_1_OF_1\",\n \"privacy_tests\": \"PASS_18_OF_18\",\n \"brand_contract\": \"PASS_19_OF_19_PLUS_VERIFIER\",\n \"diff_check\": \"PASS\",\n \"codeql\": \"PASS_NEW_HEAD\",\n \"vercel_deployment\": \"READY_SOURCE_BOUND\",\n \"vercel_deployment_id\": \"dpl_DPREtpZhVSwkwY2e3Po29FYk68UH\",\n \"vercel_deployment_sha\": \"08e0457569d734518569133709d97592e320fd42\",\n \"preview_support_semantics\": \"PASS\",\n \"preview_privacy_semantics\": \"PASS\",\n \"preview_terms_semantics\": \"PASS\",\n \"preview_body_readback\": \"PASS_ROOT_SHARE_COOKIE_FLOW\",\n \"live_semantics\": \"FAIL_NOT_PROMOTED\",\n \"supersedes_after_review\": [83, 88]\n },\n \"scan_tools\": {\n \"status\": \"NOT_EXECUTED\",\n \"snapshot\": null\n },\n \"submission\": {\n \"submit_for_review\": \"NOT_EXECUTED\",\n \"publication\": \"NOT_EXECUTED\"\n },\n \"invariant\": \"TECHNICAL_PASS != PORTAL_ACCESS != PRODUCTION_HOST_AUTHORITY != PORTAL_METADATA_COMPLETE != SCAN_TOOLS_PASS != SUBMISSION_AUTHORITY != PUBLICATION\"\n}",[225],"ab81246c-c07a-4d2b-93e4-b13b5b6d2164","PORTAL_LOGO.svg","bc6341dc-ada3-11f1-95b6-1607ee4eb77e",49,["D",1789105417674],"\n",[232],"e55f6bbc-0bce-45bd-80c9-01c797f79923","PORTAL_EXECUTION_RUNBOOK.md","eb037be6-ada4-11f1-93c1-1607ee4eb77e",67,["D",1789110699747],1,"# RUMBO IA CRM — OpenAI Plugin Submission Portal Runbook\n\nDate: 2026-09-11 UTC\nBudget invariant: `MONEY_SPEND=0`\nSubmission authority: `NOT_GRANTED`\nPublication authority: `NOT_GRANTED`\n\nSource of truth:\n- Candidate SHA: `9152f5a50de42be55508f0effed5ed09aee44c60`\n- OpenAI organization: `Rumbo`\n- Connected project: `Default project`\n- Submission portal: `https://platform.openai.com/plugins`\n- Submission type: `With MCP`\n- MCP URL type: `Universal`\n\n## Gate 0 — do not start submission until these are proven\n\n1. The acting account can access the current OpenAI app-creation/submission surface for its plan, workspace, and role. For workspace custom apps, current OpenAI documentation requires an eligible Business or Enterprise/Edu workspace and the corresponding admin/authorized-developer role; public directory submission is a separate flow.\n2. Any publisher/developer identity field actually required by the live submission portal is resolved and matches the public RUMBO IA listing. Do not invent an identity requirement that the portal does not ask for.\n3. Project data residency is informational unless the live portal or current official submission documentation presents it as a blocking requirement. Do not block solely on community reports.\n4. A production MCP URL has been explicitly authorized. The current `rumbo-crm-openai-staging.val.run` endpoint remains a technically verified staging/review candidate and must not be represented as production without separate authority.\n\n## Info tab\n\nUse `PORTAL_PACKET.json` as the source of truth.\n\n- Name: `RUMBO IA CRM`\n- Category: `PRODUCTIVITY`\n- Short/long descriptions: copy from `PORTAL_PACKET.json`.\n- Website: `https://rumbo-crm-openai-staging.val.run/` for the self-contained review packet unless publisher policy requires the corporate website field; do not substitute a stale policy page.\n- Support: `https://rumbo-crm-openai-staging.val.run/support`\n- Privacy: `https://rumbo-crm-openai-staging.val.run/privacy`\n- Terms: `https://rumbo-crm-openai-staging.val.run/terms`\n- Logo asset: `PORTAL_LOGO.svg`, SHA-256 `3ac48eb17d735ee41d12e6edf0cc0ea0e5a55358cce6b5fd37d93b3fc0372205`.\n- Developer Identity: select only a verified identity that matches the public listing.\n\n## MCP tab\n\n1. Select `Universal`.\n2. Enter the **authorized production** MCP URL. Do not submit the current staging hostname as production until that promotion is explicitly authorized.\n3. Authentication: `None` for the current bounded public v1.\n4. Custom UI: none; CSP is not applicable because this v1 exposes no UI resources.\n5. If the portal displays `Domain not verified`, copy the exact generated token into the Val Town environment variable `OPENAI_APPS_CHALLENGE_TOKEN` without exposing it in chat or source control.\n6. Verify the challenge returns only the exact token at `/.well-known/openai-apps-challenge`.\n7. Select `Scan Tools`.\n8. Require the scan to discover exactly these four tools:\n - `crm_get_service_readiness`\n - `crm_prepare_note`\n - `crm_prepare_lead_stage_change`\n - `crm_prepare_outreach_draft`\n9. Verify each scanned tool advertises `readOnlyHint=true`, `openWorldHint=false`, `destructiveHint=false`, accurate input/output schemas, and no unexpected tools/resources.\n10. If Scan Tools reports drift, stop, fix the server, rerun local gates, and scan again. Do not override a server-side annotation mismatch with portal prose.\n\n## Prompts and tests\n\n- Starter prompts: copy the four prompts from `PORTAL_PACKET.json`.\n- Submission import artifact: `chatgpt-app-submission.json` (`READY_VALIDATED_NOT_UPLOADED`). If the portal offers JSON import, upload it; otherwise copy its fields manually.\n- Positive tests: exactly 5.\n- Negative tests: exactly 3.\n- Review checks: `SUBMISSION_REVIEW_CHECKS.md`.\n- No reviewer credentials are required because authentication is `None`.\n- Do not provide UI screenshots because this v1 has no custom UI.\n\n## Global availability\n\nSelect only countries/regions that are explicitly approved for launch and for which product support and legal terms are ready. Current state is `UNSET_FAIL_CLOSED`; do not infer a distribution scope automatically.\n\n## Release notes\n\nUse the initial-submission release note from `PORTAL_PACKET.json`. Confirm that it states the bounded proposal-only surface, no customer-data reads, no CRM mutations, no autonomous sends, and no authentication requirement.\n\n## Hard stop before review\n\nBefore selecting `Submit for Review`, rerun:\n\n- `verify.ts`\n- `verify_submission_import.ts`\n- `submission_gate.ts`\n\nProceed only when `submission_gate.ts` reports no technical or portal-state blockers and separate submission authority has been granted.\n\n`TECHNICAL_PASS != PORTAL_PASS != SUBMISSION_AUTHORITY != PUBLICATION`\n\nAfter approval, publication is a separate action and remains unauthorized until explicitly granted.\n",[239],"ac45c388-45ee-4555-b082-9e697e037f49","PORTAL_EXECUTION_HANDOFF.md","deab7d64-ad5c-11f1-a2a7-1607ee4eb77e",69,["D",1789110745244],"# RUMBO IA CRM — OpenAI Apps Management execution handoff\n\nDate: 2026-09-10 UTC\nBudget: USD 0\nCandidate SHA: `9152f5a50de42be55508f0effed5ed09aee44c60`\nStatus: PRE-SUBMISSION HANDOFF ONLY — DO NOT SUBMIT OR PUBLISH\n\n## Entry conditions\n\nBefore entering data, confirm in OpenAI Platform/Apps Management:\n\n1. The acting account can access the current OpenAI app creation/submission surface for its plan, workspace, and role. For workspace custom apps, current OpenAI documentation requires eligible Business or Enterprise/Edu access; public directory submission is a separate flow.\n2. Any publisher/developer identity field actually required by the live portal is resolved and matches the public RUMBO IA listing.\n3. Project data residency is recorded if the portal exposes it, but is not treated as a blocker unless current official OpenAI submission documentation or the live portal makes it one.\n4. Live support, privacy, and terms surfaces pass `submission_gate.ts`. Current Val Town surfaces are PASS and use the same candidate SHA.\n5. The canonical RUMBO IA logo asset is hash-verified; portal upload remains unexecuted.\n6. A production MCP URL/origin has been explicitly authorized. Do not submit the current URL merely because the staging adapter passes conformance.\n\n## Draft configuration\n\nSubmission type: remote MCP-only / With MCP\nMCP URL mode: Universal\nAuthentication: None\nCustom UI: None\nReviewer credentials: Not applicable for bounded public v1\n\nUse the exact listing and release-note fields from `PORTAL_PACKET.json` after its fail-closed fields have been resolved.\nUse the 5 positive and 3 negative reviewer cases from `SUBMISSION_TESTS.md`.\n\n## Domain verification\n\nWhen Apps Management supplies the exact domain challenge token:\n\n1. Add it to the Val Town project as `OPENAI_APPS_CHALLENGE_TOKEN` using the secret/environment-variable mechanism; do not place it in source or chat logs.\n2. Verify `GET /.well-known/openai-apps-challenge` returns HTTP 200 with **only** the exact token as plaintext.\n3. Run `verify.ts` and `submission_gate.ts` again.\n4. Do not continue if either fails.\n\n## Scan Tools\n\nRun **Scan Tools** in Apps Management against the authorized production MCP URL.\nReview every discovered tool and confirm the catalog contains exactly:\n\n- `crm_get_service_readiness`\n- `crm_prepare_note`\n- `crm_prepare_lead_stage_change`\n- `crm_prepare_outreach_draft`\n\nFor each discovered tool confirm:\n\n- `readOnlyHint=true`\n- `openWorldHint=false`\n- `destructiveHint=false`\n- schema matches the bounded contract\n- no unplanned customer-data or mutating tool appears\n- no unnecessary personal data, secret, debug field, or internal identifier is returned\n\nIf the scanned catalog differs, stop and reconcile source/host before continuing.\n\n## Availability and review packet\n\nSelect only countries/regions where the publisher, product, support and legal terms are actually ready. Do not infer worldwide availability from technical reachability.\n\nEnter the starter prompts and release notes from `PORTAL_PACKET.json` and the reviewer test cases from `SUBMISSION_TESTS.md`.\n\n## Final pre-submit return gate\n\nBefore pressing **Submit for Review**, return the portal evidence to the control plane and update these fields:\n\n- `PORTAL_APP_CREATION_OR_SUBMISSION_ACCESS=PASS`\n- `PUBLISHER_IDENTITY=PASS_IF_PORTAL_REQUIRES`\n- `PROJECT_DATA_RESIDENCY=RECORDED_INFORMATIONAL_UNLESS_PORTAL_BLOCKS`\n- `LOGO_ASSET=PASS`\n- `LOGO_PORTAL_UPLOAD=PASS_IF_PORTAL_REQUIRES`\n- `SUPPORT_MCP_SEMANTICS=PASS`\n- `PRIVACY_MCP_DISCLOSURE=PASS`\n- `TERMS_BOUNDED_SEMANTICS=PASS`\n- `DOMAIN_VERIFICATION=PASS_IF_PORTAL_REQUIRES`\n- `SCAN_TOOLS=PASS`\n- `DISTRIBUTION_SCOPE=SET`\n- `PRODUCTION_MCP_URL=AUTHORIZED_AND_PROVEN`\n\nThen rerun `submission_gate.ts`.\n\nOnly `RUMBO_OPENAI_SUBMISSION_GATE_PASS` may establish technical submission readiness. It still does **not** grant authority to submit or publish.\n\n`SUBMISSION_GATE_PASS != SUBMISSION_AUTHORITY != REVIEW_APPROVAL != PUBLICATION_AUTHORITY`\n",[247],"8fcac396-d996-48e0-a487-56b955fabd00","PORTAL_EVIDENCE_RECEIPT_V2.json","d041d8aa-adb0-11f1-aa3b-1607ee4eb77e",76,["D",1789111034479],"{\n \"schema\": \"RUMBO_OPENAI_PORTAL_EVIDENCE_RECEIPT_V2\",\n \"date_utc\": \"2026-09-11\",\n \"budget_usd\": 0,\n \"candidate_sha\": \"9152f5a50de42be55508f0effed5ed09aee44c60\",\n \"technical\": {\n \"health_exact_sha\": \"PASS\",\n \"tool_annotations\": \"PASS\",\n \"tool_count\": 4,\n \"output_schemas\": \"PASS_ALL_FOUR\",\n \"listing_urls_public\": \"PASS\",\n \"support_mcp_semantics\": \"PASS\",\n \"privacy_mcp_disclosure\": \"PASS\",\n \"terms_bounded_semantics\": \"PASS\",\n \"logo_exact_asset\": \"PASS_SHA256_3ac48eb17d735ee41d12e6edf0cc0ea0e5a55358cce6b5fd37d93b3fc0372205\",\n \"submission_import\": \"PASS_4_TOOLS_5_POSITIVE_3_NEGATIVE\"\n },\n \"official_product_model\": {\n \"workspace_custom_mcp\": \"BUSINESS_OR_ENTERPRISE_EDU_WITH_ROLE_CONTROLS\",\n \"workspace_creation_flow\": \"ENDPOINT_METADATA_THEN_SCAN_TOOLS_THEN_CREATE\",\n \"public_directory_submission\": \"SEPARATE_FLOW\",\n \"plugin_directory\": \"PRIMARY_DISCOVERY_SURFACE_SINCE_2026_07_09\",\n \"global_data_residency_universal_submission_blocker\": \"NOT_PROVEN_IN_CURRENT_OFFICIAL_SOURCES\",\n \"apps_management_write_exact_permission_universal_requirement\": \"NOT_PROVEN_IN_CURRENT_OFFICIAL_SOURCES\"\n },\n \"portal_observation\": {\n \"status\": \"UNVERIFIED_BY_AUTOMATION\",\n \"desktop\": \"DESKTOP-QUGVQLB_ONLINE\",\n \"cdp_127_0_0_1_9222\": \"NO_LISTENER\",\n \"github_live_reconciliation\": \"UNAVAILABLE_CONNECTOR_DISABLED\"\n },\n \"challenge\": {\n \"capability\": \"IMPLEMENTED_FAIL_CLOSED\",\n \"without_token\": \"HTTP_404\",\n \"verification\": \"NOT_EXECUTED\"\n },\n \"current_blockers\": [\n \"domain_verification_not_completed\",\n \"portal_app_creation_or_submission_access_unverified\",\n \"publisher_identity_portal_requirement_unresolved\",\n \"logo_portal_upload_not_executed\",\n \"scan_tools_not_executed\",\n \"distribution_scope_unset\",\n \"production_mcp_url_not_authorized\"\n ],\n \"authority\": {\n \"production\": \"NOT_GRANTED\",\n \"submit_for_review\": \"NOT_GRANTED_AND_NOT_EXECUTED\",\n \"publication\": \"NOT_GRANTED_AND_NOT_EXECUTED\"\n },\n \"invariant\": \"TECHNICAL_PASS != PORTAL_ACCESS != PRODUCTION_HOST_AUTHORITY != PORTAL_METADATA_COMPLETE != SCAN_TOOLS_PASS != SUBMISSION_AUTHORITY != PUBLICATION\"\n}",[254],"pulls","branches","forks","envVars",{"_267":63,"_269":63},"connections",{"_267":63,"_268":63},"count","unfilledCount","emptyCount",{"_43":253,"_32":254,"_30":255,"_147":-5,"_41":42,"_40":256,"_149":59,"_38":332,"_13":63,"_33":20,"_35":258,"_24":25,"_152":-5,"_287":333,"_27":334},{"_43":246,"_32":247,"_30":248,"_147":-5,"_41":42,"_40":249,"_149":25,"_38":329,"_13":207,"_33":20,"_35":251,"_24":25,"_152":-5,"_287":330,"_27":331},{"_43":238,"_32":239,"_30":240,"_147":-5,"_41":42,"_40":241,"_149":25,"_38":326,"_13":243,"_33":20,"_35":244,"_24":25,"_152":-5,"_287":327,"_27":328},{"_43":231,"_32":232,"_30":233,"_147":-5,"_41":42,"_40":234,"_149":59,"_38":323,"_13":63,"_33":20,"_35":236,"_24":25,"_152":-5,"_287":324,"_27":325},{"_43":224,"_32":225,"_30":226,"_147":-5,"_41":42,"_40":227,"_149":25,"_38":320,"_13":163,"_33":20,"_35":229,"_24":25,"_152":-5,"_287":321,"_27":322},{"_43":217,"_32":218,"_30":219,"_147":-5,"_41":42,"_40":220,"_149":59,"_38":317,"_13":63,"_33":20,"_35":222,"_24":25,"_152":-5,"_287":318,"_27":319},{"_43":210,"_32":211,"_30":212,"_147":-5,"_41":42,"_40":213,"_149":59,"_38":314,"_13":63,"_33":20,"_35":215,"_24":25,"_152":-5,"_287":315,"_27":316},{"_43":202,"_32":203,"_30":204,"_147":-5,"_41":42,"_40":205,"_149":59,"_38":311,"_13":207,"_33":20,"_35":208,"_24":25,"_152":-5,"_287":312,"_27":313},{"_43":195,"_32":196,"_30":197,"_147":-5,"_41":42,"_40":198,"_149":25,"_38":308,"_13":184,"_33":20,"_35":200,"_24":25,"_152":-5,"_287":309,"_27":310},{"_43":188,"_32":189,"_30":190,"_147":-5,"_41":42,"_40":191,"_149":25,"_38":305,"_13":63,"_33":20,"_35":193,"_24":25,"_152":-5,"_287":306,"_27":307},{"_43":181,"_32":182,"_30":183,"_147":-5,"_41":42,"_40":184,"_149":59,"_38":302,"_13":63,"_33":20,"_35":186,"_24":25,"_152":-5,"_287":303,"_27":304},{"_43":174,"_32":175,"_30":176,"_147":-5,"_41":42,"_40":177,"_149":59,"_38":299,"_13":63,"_33":20,"_35":179,"_24":25,"_152":-5,"_287":300,"_27":301},{"_43":166,"_32":167,"_30":57,"_147":-5,"_41":42,"_40":168,"_149":25,"_38":296,"_13":170,"_33":171,"_35":172,"_24":25,"_152":45,"_287":297,"_27":298},{"_43":158,"_32":159,"_30":160,"_147":-5,"_41":42,"_40":161,"_149":25,"_38":293,"_13":163,"_33":34,"_35":164,"_24":25,"_152":-5,"_287":294,"_27":295},{"_43":154,"_32":28,"_30":31,"_147":-5,"_41":42,"_40":14,"_149":25,"_38":290,"_13":156,"_33":34,"_35":37,"_24":25,"_152":-5,"_287":291,"_27":292},{"_43":144,"_32":145,"_30":146,"_147":-5,"_41":42,"_40":148,"_149":59,"_38":286,"_13":63,"_33":34,"_35":151,"_24":25,"_152":-5,"_287":288,"_27":289},["D",1789105772539],"children",[],[145],["D",1789115828327],[],[28],["D",1789110934250],[],[159],["D",1789105438225],[],[167],["D",1789105584348],[],[175],["D",1789072894892],[],[182],["D",1789105639099],[],[189],["D",1789110984832],[],[196],["D",1789072867612],[],[203],["D",1789102080178],[],[211],["D",1789074813327],[],[218],["D",1789111012858],[],[225],["D",1789105417674],[],[232],["D",1789110699747],[],[239],["D",1789110745244],[],[247],["D",1789111034479],[],[254],["D",1789031650837],{"_97":96,"_105":101},{"_97":96,"_98":99,"_100":101,"_18":55,"_102":103,"_104":25},{"_13":86,"_32":87,"_43":42,"_88":44,"_89":-5,"_90":-5,"_91":-5,"_92":-5,"_93":-5,"_94":45,"_95":96},"ctx",{"_79":-5,"_125":372,"_373":374,"_375":376,"_377":25,"_378":25,"_379":25,"_380":59,"_381":-5,"_382":-7,"_383":-7,"_384":63,"_385":25,"_386":25,"_387":-5,"_388":25,"_389":25,"_390":25,"_391":-5},"ENV",{"_343":344,"_345":346,"_347":87,"_348":349,"_350":351,"_352":353,"_354":355,"_356":357,"_358":359,"_360":361,"_362":363,"_364":365,"_366":367,"_368":369,"_370":371},"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"]

Diff · v78