n8n 2.37.5 → 2.38.0

package/dist/modules/agents/builder/prompts/llm-selection.prompt.js

@@ -15,12 +15,13 @@ Use this to resolve the target agent's main \`model\` and \`credential\`.
 ### Workflow
 
 1. For fresh agents, call \`read_config\` first. If \`model\` and \`credential\` are already set (the system auto-selected a sensible default at creation), keep them and mention the choice as changeable in your summary — do not call \`resolve_llm\`. Otherwise call \`resolve_llm\` once, silently, before the first config write — with provider/model when the user named them, otherwise with no arguments.
-2. If \`resolve_llm\` succeeds, persist \`model = "{provider}/{model}"\` and \`credential = credentialId\`. If the result has \`claimedFreeOpenAiCredits: true\`, tell the user you set them up with free OpenAI credits. If the result's \`credentialName\` is "n8n credits" (its \`credentialId\` is the managed tag), persist it like any credential and tell the user the model runs on n8n credits — never surface the internal "n8n Connect" or "AI Gateway" names. If it has \`autoPicked: true\`, tell the user which provider and model you picked and that they can ask to change it — do not ask for confirmation and do not raise a trailing model question.
+2. If \`resolve_llm\` succeeds, persist \`model = "{provider}/{model}"\` and \`credential = credentialId\`. If the result has \`claimedFreeOpenAiCredits: true\`, tell the user you set them up with free OpenAI credits. If the result's \`credentialName\` is "Gateway credits" (its \`credentialId\` is the managed tag), persist it like any credential and tell the user the model runs on Gateway credits — never surface the internal "n8n Connect" or "AI Gateway" names. If it has \`autoPicked: true\`, tell the user which provider and model you picked and that they can ask to change it — do not ask for confirmation and do not raise a trailing model question.
 3. If the user asks to pick, change, confirm, or configure a model or main credential, ask via \`ask_questions\`; do not ask in prose.
-   - Exception — when the user explicitly asks to use n8n credits for the main model, do NOT ask: call \`resolve_llm\` with \`useN8nCredits: true\` (and \`provider\` when the user named one). It resolves n8n credits for that provider even if the user already has their own credential for it; persist the result (\`credentialName\` "n8n credits") and tell the user the model runs on n8n credits. If it returns \`n8n_credits_unsupported_provider\`, tell the user n8n credits does not cover that provider and offer their own credential; if \`ambiguous_n8n_credits_provider\`, ask which provider via \`ask_questions\` using the returned \`providers\`; if \`n8n_credits_unavailable\`, tell the user n8n credits is not available on this instance and resolve their own credential instead. On \`unknown_model\`, retry with a value from the returned \`availableModels\` or surface those options.
+   - Exception — when the user explicitly asks to use Gateway credits for the main model, do NOT ask: call \`resolve_llm\` with \`useGatewayCredits: true\` (and \`provider\` when the user named one). It resolves Gateway credits for that provider even if the user already has their own credential for it; persist the result (\`credentialName\` "Gateway credits") and tell the user the model runs on Gateway credits. If it returns \`gateway_credits_unsupported_provider\`, tell the user Gateway credits do not cover that provider and offer their own credential; if \`ambiguous_gateway_credits_provider\`, ask which provider via \`ask_questions\` using the returned \`providers\`; if \`gateway_credits_unavailable\`, tell the user Gateway credits are not available on this instance and resolve their own credential instead. On \`unknown_model\`, retry with a value from the returned \`availableModels\` or surface those options.
 4. During an initial build, if \`resolve_llm\` reports missing or ambiguous credentials/provider, do not ask: mark the model task \`blocked\`, keep building with \`model: ""\` and no \`credential\`, and include the model choice as a question in the trailing \`finish_setup\` call — for ambiguity between multiple credentials of one provider, use the credential names from the resolve_llm result as the question's options — when it resolves, call \`resolve_llm\` with the answer (pass \`credentialId\` when the user picked a specific credential) and patch \`/model\` and \`/credential\` — after \`read_config\`, since the user may have already set the model in the panel. For a model change on an existing agent, ask immediately instead, and never write \`model: ""\` over an existing model — keep the current model and credential until the new one is resolved.
-5. If \`resolve_llm\` reports \`unknown_model\`, retry with a plausible returned model value or ask via \`ask_questions\`. If it reports \`model_lookup_failed\` (the provider's model list could not be fetched, e.g. a transient error — possible even on the default no-model path), retry \`resolve_llm\`; do not guess a model.
-6. If the model is still unresolved when the user asks to run or publish the agent, leave the draft with \`model: ""\`, do not guess a model, and tell the user the agent needs a model and credential first — in the panel or here in chat. If they dismiss a model-change question on an existing agent, keep the current model and credential unchanged.
+5. If \`resolve_llm\` reports \`unknown_model\`, retry with a plausible value from the returned \`availableModels\` or ask via \`ask_questions\`. This can happen on the default no-model path too, when the credential cannot reach the provider's default model — the returned \`availableModels\` is that credential's live list, so pick from it rather than guessing or reusing the id you just tried. If it reports \`model_lookup_failed\` (the provider's model list could not be fetched, e.g. a transient error — possible even on the default no-model path), retry \`resolve_llm\`; do not guess a model.
+6. If \`call_agent\` fails with \`code: "invalid_model"\`, the provider rejected the model id, not the credential. Do not tell the user to check, revoke, or regenerate their API key. Call \`resolve_llm\` for that provider again, persist a model from its result, and retry the test run.
+7. If the model is still unresolved when the user asks to run or publish the agent, leave the draft with \`model: ""\`, do not guess a model, and tell the user the agent needs a model and credential first — in the panel or here in chat. If they dismiss a model-change question on an existing agent, keep the current model and credential unchanged.
 
 ### Rules
 

package/dist/modules/agents/builder/prompts/llm-selection.prompt.js.map

@@ -1 +1 @@
-{"version":3,"file":"llm-selection.prompt.js","sourceRoot":"","sources":["../../../../../src/modules/agents/builder/prompts/llm-selection.prompt.ts"],"names":[],"mappings":";;;AAAA,+BAAsC,2BAA0C;IAC/E,MAAM,sBAAsB,GAAG,2BAA2B;QACzD,CAAC,CAAC,OAAO,2BAA2B,EAAE;QACtC,CAAC,CAAC,8MAA8M,CAAC;IAElN,OAAO;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;iJAqDyI,sBAAsB,EAAE,CAAC;AAC1K,CAAC"}
+{"version":3,"file":"llm-selection.prompt.js","sourceRoot":"","sources":["../../../../../src/modules/agents/builder/prompts/llm-selection.prompt.ts"],"names":[],"mappings":";;;AAAA,+BAAsC,2BAA0C;IAC/E,MAAM,sBAAsB,GAAG,2BAA2B;QACzD,CAAC,CAAC,OAAO,2BAA2B,EAAE;QACtC,CAAC,CAAC,8MAA8M,CAAC;IAElN,OAAO;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;iJAsDyI,sBAAsB,EAAE,CAAC;AAC1K,CAAC"}

package/dist/modules/agents/builder/prompts/tools.prompt.d.ts

@@ -1 +1 @@
-export declare const TOOLS_PROMPT = "## Tool Guidance\n\n### Purpose\n\nUse this to give the target agent callable capabilities through MCP servers,\nworkflows, nodes, custom code tools, or provider tools.\n\n### Workflow\n\nUse this guidance before calling `resolve_integration`, `search_nodes`,\n`search_mcp_servers`, `get_node_types`, `build_custom_tool`, or adding,\nchanging, or removing entries in `tools[]` / `mcpServers` / `providerTools`.\n\nFor an external product, load `agent-builder-external-services` once and\nfollow it. It covers the chat-integration-versus-callable-tool decision, chat\nintegration setup, and MCP servers. Load `agent-builder-node-tools` before\nconfiguring a node tool.\n\n- Chat/trigger integration: call `list_integration_types`, then\n  `configure_channel` with a returned type. Do not call `resolve_integration`\n  for chat/trigger integrations.\n- Callable external service: the conversation or trigger happens elsewhere and\n  the agent only operates on the product. For each requested non-chat callable\n  service, call `resolve_integration` separately, using `queries` as\n  alternative search terms for that one service. Do not infer MCP availability\n  from memory.\n  - `kind: \"mcp\"`: follow the skill's MCP Servers section \u2014 credential,\n    verification, and config workflow.\n  - `kind: \"node\"`: load `agent-builder-node-tools`, use the returned node\n    results, and continue with `get_node_types`.\n\nUse `search_nodes` directly only when the user explicitly asks for an n8n node,\nwhen refining node results, or when a verified MCP server lacks the requested\ncapability. Use `search_mcp_servers` directly only when refining MCP results or\nwhen the user explicitly asks to browse the MCP registry.\n\nPreference order for non-chat callable services:\n1. MCP servers selected by `resolve_integration`\n2. Node tools returned by `resolve_integration`\n3. Workflow tools (`list_workflows`)\n4. Custom tools (`build_custom_tool`) \u2014 last resort\n\nCustom tools are for pure computation, validation, formatting, or planning logic;\nthey cannot perform live network, filesystem, process, timer, or host I/O.\n\n#### Workflow Tools\n\n- Call `list_workflows`; reference supported workflows with `{ \"type\": \"workflow\", \"workflowId\": \"<id>\", \"workflow\": \"<name>\" }` using both values it returns.\n\n#### Node Tools\n\nLoad `agent-builder-node-tools` before adding, changing, or removing\nnode-backed tools, `nodeParameters`, `$fromAI` usage, or n8n expressions.\nFor an HTTP Request Tool, use only an exact URL explicitly supplied by the\nuser. If the user has not supplied one during an initial build, you MUST ask\nfor it through the trailing `finish_setup` call, then configure the tool with\nthe answer. On later turns, use `ask_questions` before mutating the config.\nNever infer or invent a URL.\n\n#### MCP Servers\n\nLoad `agent-builder-external-services` when the user explicitly requests a\ncustom MCP server, and follow its MCP Servers section.\n\n#### Custom Tools\n\nCustom tools are last resort and only for pure computation. Load\n`agent-builder-custom-tools` before calling `build_custom_tool`.\n\n#### Provider Tools\n\n- Match provider tools to the configured model provider.\n- Anthropic: `providerTools[\"anthropic.web_search\"]`.\n- OpenAI: `providerTools[\"openai.web_search\"]` or `providerTools[\"openai.image_generation\"]`, only for compatible OpenAI models.\n\n### Gotchas\n\n- Generic web search must use `config.webSearch`; never add an HTTP Request\n  Tool unless the user explicitly requests direct HTTP, API, or specific-page\n  fetching.\n- Live crawling, fetching, and API integrations need MCP, workflow, or node tools, not custom tools.\n- Do not invent MCP servers, node type names, workflow names, credential ids, or provider tool keys.\n\n### Verify\n\n- Chat/trigger integrations bypassed `resolve_integration` and were set up\n  through `configure_channel`.\n- Each non-chat callable service was resolved separately through\n  `resolve_integration` unless the user explicitly requested an n8n node or\n  custom MCP server.\n- Workflow tools reference discovered workflow IDs and names.\n- Provider tool keys match the configured model provider and the valid key list.";
+export declare const TOOLS_PROMPT = "## Tool Guidance\n\n### Purpose\n\nUse this to give the target agent callable capabilities through MCP servers,\nworkflows, nodes, custom code tools, or provider tools.\n\n### Workflow\n\nUse this guidance before calling `resolve_integration`, `search_nodes`,\n`search_mcp_servers`, `get_node_types`, `build_custom_tool`, or adding,\nchanging, or removing entries in `tools[]` / `mcpServers` / `providerTools`.\n\nFor an external product, load `agent-builder-external-services` once and\nfollow it. It covers the chat-integration-versus-callable-tool decision, chat\nintegration setup, and MCP servers. Load `agent-builder-node-tools` before\nconfiguring a node tool.\n\n- Chat/trigger integration: call `list_integration_types`, then\n  `configure_channel` with a returned type. Do not call `resolve_integration`\n  for chat/trigger integrations.\n- A configured chat integration generates its own context and action tools for\n  every top-level Agent run, including scheduled tasks. When its capabilities\n  include the requested action, do not add a same-platform node, MCP, or\n  workflow tool. A scheduled or proactive message with no inbound conversation\n  is not, by itself, a reason to add another tool.\n- Callable external service: the conversation or trigger happens elsewhere and\n  the agent only operates on the product. For each requested non-chat callable\n  service, call `resolve_integration` separately, using `queries` as\n  alternative search terms for that one service. Do not infer MCP availability\n  from memory.\n  - `kind: \"mcp\"`: follow the skill's MCP Servers section \u2014 credential,\n    verification, and config workflow.\n  - `kind: \"node\"`: load `agent-builder-node-tools`, use the returned node\n    results, and continue with `get_node_types`.\n\nUse `search_nodes` directly only when the user explicitly asks for an n8n node,\nwhen refining node results, or when a verified MCP server lacks the requested\ncapability. Use `search_mcp_servers` directly only when refining MCP results or\nwhen the user explicitly asks to browse the MCP registry.\n\nPreference order for non-chat callable services:\n1. MCP servers selected by `resolve_integration`\n2. Node tools returned by `resolve_integration`\n3. Workflow tools (`list_workflows`)\n4. Custom tools (`build_custom_tool`) \u2014 last resort\n\nCustom tools are for pure computation, validation, formatting, or planning logic;\nthey cannot perform live network, filesystem, process, timer, or host I/O.\n\n#### Workflow Tools\n\n- Call `list_workflows`; reference supported workflows with `{ \"type\": \"workflow\", \"workflowId\": \"<id>\", \"workflow\": \"<name>\" }` using both values it returns.\n\n#### Node Tools\n\nLoad `agent-builder-node-tools` before adding, changing, or removing\nnode-backed tools, `nodeParameters`, `$fromAI` usage, or n8n expressions.\nFor an HTTP Request Tool, use only an exact URL explicitly supplied by the\nuser. If the user has not supplied one during an initial build, you MUST ask\nfor it through the trailing `finish_setup` call, then configure the tool with\nthe answer. On later turns, use `ask_questions` before mutating the config.\nNever infer or invent a URL.\n\n#### MCP Servers\n\nLoad `agent-builder-external-services` when the user explicitly requests a\ncustom MCP server, and follow its MCP Servers section.\n\n#### Custom Tools\n\nCustom tools are last resort and only for pure computation. Load\n`agent-builder-custom-tools` before calling `build_custom_tool`.\n\n#### Provider Tools\n\n- Match provider tools to the configured model provider.\n- Anthropic: `providerTools[\"anthropic.web_search\"]`.\n- OpenAI: `providerTools[\"openai.web_search\"]` or `providerTools[\"openai.image_generation\"]`, only for compatible OpenAI models.\n\n### Gotchas\n\n- Generic web search must use `config.webSearch`; never add an HTTP Request\n  Tool unless the user explicitly requests direct HTTP, API, or specific-page\n  fetching.\n- Live crawling, fetching, and API integrations need MCP, workflow, or node tools, not custom tools.\n- Do not invent MCP servers, node type names, workflow names, credential ids, or provider tool keys.\n- Do not duplicate an action already supplied by a configured chat integration.\n\n### Verify\n\n- Chat/trigger integrations bypassed `resolve_integration` and were set up\n  through `configure_channel`.\n- Each non-chat callable service was resolved separately through\n  `resolve_integration` unless the user explicitly requested an n8n node or\n  custom MCP server.\n- Workflow tools reference discovered workflow IDs and names.\n- Provider tool keys match the configured model provider and the valid key list.";

package/dist/modules/agents/builder/prompts/tools.prompt.js

@@ -23,6 +23,11 @@ configuring a node tool.
 - Chat/trigger integration: call \`list_integration_types\`, then
   \`configure_channel\` with a returned type. Do not call \`resolve_integration\`
   for chat/trigger integrations.
+- A configured chat integration generates its own context and action tools for
+  every top-level Agent run, including scheduled tasks. When its capabilities
+  include the requested action, do not add a same-platform node, MCP, or
+  workflow tool. A scheduled or proactive message with no inbound conversation
+  is not, by itself, a reason to add another tool.
 - Callable external service: the conversation or trigger happens elsewhere and
   the agent only operates on the product. For each requested non-chat callable
   service, call \`resolve_integration\` separately, using \`queries\` as
@@ -84,6 +89,7 @@ Custom tools are last resort and only for pure computation. Load
   fetching.
 - Live crawling, fetching, and API integrations need MCP, workflow, or node tools, not custom tools.
 - Do not invent MCP servers, node type names, workflow names, credential ids, or provider tool keys.
+- Do not duplicate an action already supplied by a configured chat integration.
 
 ### Verify
 

package/dist/modules/agents/builder/prompts/tools.prompt.js.map

@@ -1 +1 @@
-{"version":3,"file":"tools.prompt.js","sourceRoot":"","sources":["../../../../../src/modules/agents/builder/prompts/tools.prompt.ts"],"names":[],"mappings":";;;AAAa,QAAA,YAAY,GAAG;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;iFA4FqD,CAAC"}
+{"version":3,"file":"tools.prompt.js","sourceRoot":"","sources":["../../../../../src/modules/agents/builder/prompts/tools.prompt.ts"],"names":[],"mappings":";;;AAAa,QAAA,YAAY,GAAG;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;iFAkGqD,CAAC"}

package/dist/modules/agents/builder/skills/external-services.skill.js

@@ -31,6 +31,7 @@ function externalServicesSkill() {
             'read_config',
             'patch_config',
             'write_config',
+            'report_required_artifact',
             'load_skill',
         ],
         instructions: `\
@@ -47,15 +48,28 @@ Use an integration when the product is the agent's conversation or trigger
 surface: humans will mention, message, comment to, or resume the agent there,
 or the agent needs to respond in that same platform conversation context.
 
+Native Agent chat integrations are bidirectional within their conversation
+context. The integration both receives the triggering message and delivers the
+Agent's replies. Do not add a same-platform node, MCP, or workflow tool merely
+so the Agent can reply, send a normal conversational message, or use interaction
+features already listed in that integration's capabilities.
+
+Configured integrations also generate their listed context and action tools for
+every top-level Agent run, including scheduled tasks. A scheduled task can use
+actions such as \`send_dm\` or \`send_channel_message\` without an inbound
+message. Being proactive, scheduled, or outside an open conversation is never
+by itself a reason to add a same-platform node, MCP, or workflow tool.
+
 Use an MCP, node, or workflow tool when the product is only something the agent
 operates on: searching records, creating tickets, updating objects, or sending a
 business-process notification while the conversation happens elsewhere.
 
-When building an agent that should interact with Slack, Discord, or Telegram,
-prefer the matching chat integration over an MCP, node, or workflow tool, even
-when a callable tool could perform the same messaging action. Use a callable
-tool instead only when the user explicitly asks for one or the requested
-operation is not supported by the chat integration.
+When building an agent that should interact with Slack, Discord, Telegram, or
+Linear, use the matching chat integration instead of an MCP, node, or workflow
+tool when the platform is the conversation surface. Add a callable tool only
+for an explicitly requested operation that is absent from the selected
+integration's returned capabilities. Compare the exact operation; do not infer
+that the integration is unavailable merely because the run starts from a task.
 
 Examples:
 
@@ -65,7 +79,8 @@ Examples:
 - Discord integration: the agent should be mentioned or messaged in Discord,
   respond in Discord threads or DMs, or render approval buttons there.
 - Telegram integration: the agent should receive or send Telegram messages,
-  continue conversations there, or render supported interactive messages.
+  continue conversations there, render supported interactive messages, or use
+  \`send_dm\` to initiate a scheduled message to a known Telegram user ID.
 - Linear integration: the agent should be triggered from Linear issues/comments,
   understand the current Linear subject, or reply in the same Linear
   conversation.
@@ -73,10 +88,35 @@ Examples:
   workflow and only needs to search/create/update Linear tickets via MCP or node
   tools.
 
+If \`list_integration_types\` does not return the requested conversation
+platform, do not substitute a platform messaging node as an Agent tool. The
+Agent needs an external channel-bridge workflow instead:
+
+1. Finish the Agent without a native integration for that platform.
+2. When \`report_required_artifact\` is available, call it once with an
+   \`artifact\` whose \`type\` is \`"workflow"\` and \`relationship\` is
+   \`"agent-entrypoint"\`. Require the
+   platform trigger, Message an Agent using the incoming message and a stable
+   platform conversation/sender identifier as its custom session key, and the
+   platform send action using the Agent's text response.
+3. The bridge invokes the Agent. Never add it to the Agent's \`tools\` array and
+   never report it as \`relationship: "agent-tool"\`.
+4. If the reporting tool is unavailable, state the same workflow requirement
+   clearly in the final reply so the calling surface can create it.
+
 For callable (non-chat) services, call \`resolve_integration\` separately per
 service and follow the returned \`kind\`: \`"mcp"\` -> MCP Servers section
 below, \`"node"\` -> load \`agent-builder-node-tools\`.
 
+### Correcting a redundant channel tool
+
+If the user questions why a same-platform tool exists, inspect the current
+config and the integration's returned capabilities before answering. When the
+integration supplies the tool's purpose, remove the redundant node, MCP, or
+workflow tool and explain the correction. Do not defend it merely because the
+message is proactive, scheduled, or has no current conversation; generated
+actions such as \`send_dm\` are available without inbound message context.
+
 ## Chat Integrations
 
 The \`integrations\` array controls how the target agent is triggered.
@@ -118,8 +158,9 @@ The \`integrations\` array controls how the target agent is triggered.
 - Do not add a chat integration just because the agent needs CRUD or notifications
   for that product. Resolve the callable capability through \`resolve_integration\`
   unless the product itself is the chat/trigger context.
-- For recurring or scheduled runs, create a task (\`create_tasks\`), not an
-  integration.
+- For recurring or scheduled runs, create a task (\`create_tasks\`) for the
+  cadence. Keep a requested chat integration, and use its generated action
+  tools when the task sends through that same platform.
 - Omitting \`integrations\` from a config write preserves the current channels.
   To remove one, write an explicit filtered array or remove the exact array
   entry.
@@ -294,6 +335,8 @@ Auth, or None) via \`${api_types_1.ASK_QUESTIONS_TOOL_NAME}\`. Then map to:
 - The chosen integration matches \`useIntegrationWhen\`; otherwise resolve the
   callable capability through \`resolve_integration\` and use MCP, node, or
   workflow tools.
+- No node, MCP, or workflow tool duplicates an action listed by a configured
+  chat integration, including for proactive scheduled tasks.
 - Generic non-chat external services were routed through \`resolve_integration\`
   before MCP or node setup.
 - The final \`integrations\` array keeps unrelated integrations intact and

package/dist/modules/agents/builder/skills/external-services.skill.js.map

@@ -1 +1 @@
-{"version":3,"file":"external-services.skill.js","sourceRoot":"","sources":["../../../../../src/modules/agents/builder/skills/external-services.skill.ts"],"names":[],"mappings":";;;AACA,8CAAgF;AAEhF,2DAAqD;AAErD,qFAAmF;AACnF,0EAAqE;AAErE,MAAM,mBAAmB,GAAG,IAAA,gDAAuB,EAClD,IAAA,oCAAe,EAAC,iCAAqB,CAAgB,CACrD,CAAC;AAEF;IACC,OAAO;QACN,EAAE,EAAE,iCAAiC;QACrC,IAAI,EAAE,iCAAiC;QACvC,WAAW,EACV,kQAAkQ;QACnQ,gBAAgB,EAAE;YACjB,qBAAqB;YACrB,wBAAwB;YACxB,mBAAmB;YACnB,gBAAgB;YAChB,mBAAmB;YACnB,aAAa;YACb,cAAc;SACd;QACD,YAAY,EAAE;YACb,qBAAqB;YACrB,wBAAwB;YACxB,mBAAmB;YACnB,oBAAoB;YACpB,gBAAgB;YAChB,mBAAmB;YACnB,eAAe;YACf,aAAa;YACb,cAAc;YACd,cAAc;YACd,YAAY;SACZ;QACD,YAAY,EAAE;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;IA0DZ,yCAAkB;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;MAuDhB,mCAAuB;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;EAgC3B,yCAAkB;;;;;;;;;;;;;;;;;;EAkBlB,mBAAmB;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;uBAiEE,mCAAuB;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;8CAoCA;KAC5C,CAAC;AACH,CAAC"}
+{"version":3,"file":"external-services.skill.js","sourceRoot":"","sources":["../../../../../src/modules/agents/builder/skills/external-services.skill.ts"],"names":[],"mappings":";;;AACA,8CAAgF;AAEhF,2DAAqD;AAErD,qFAAmF;AACnF,0EAAqE;AAErE,MAAM,mBAAmB,GAAG,IAAA,gDAAuB,EAClD,IAAA,oCAAe,EAAC,iCAAqB,CAAgB,CACrD,CAAC;AAEF;IACC,OAAO;QACN,EAAE,EAAE,iCAAiC;QACrC,IAAI,EAAE,iCAAiC;QACvC,WAAW,EACV,kQAAkQ;QACnQ,gBAAgB,EAAE;YACjB,qBAAqB;YACrB,wBAAwB;YACxB,mBAAmB;YACnB,gBAAgB;YAChB,mBAAmB;YACnB,aAAa;YACb,cAAc;SACd;QACD,YAAY,EAAE;YACb,qBAAqB;YACrB,wBAAwB;YACxB,mBAAmB;YACnB,oBAAoB;YACpB,gBAAgB;YAChB,mBAAmB;YACnB,eAAe;YACf,aAAa;YACb,cAAc;YACd,cAAc;YACd,0BAA0B;YAC1B,YAAY;SACZ;QACD,YAAY,EAAE;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;IAiGZ,yCAAkB;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;MAwDhB,mCAAuB;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;EAgC3B,yCAAkB;;;;;;;;;;;;;;;;;;EAkBlB,mBAAmB;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;uBAiEE,mCAAuB;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;8CAsCA;KAC5C,CAAC;AACH,CAAC"}

package/dist/modules/agents/builder/skills/node-tools.skill.js

@@ -80,10 +80,10 @@ conventions. Never add an incomplete tool or use a placeholder URL.
 - Never write literal \`"$fromAI"\` or bare \`$fromAI\`; the node will treat it as the actual value.
 - Do not pipe AI-chosen fields through \`$json\`.
 - Do not include \`inputSchema\` or \`toolDescription\` for node tools.
-- n8n Connect (\`n8n credits\`) covers many services, including some community nodes. Adding a node tool with its credential slot omitted triggers server-side assignment: for a covered service the server attaches the managed \`n8n credits\` credential (\`{ id: null, name: "n8n credits", __aiGatewayManaged: true }\`) to each required, eligible slot on write — but only when the project has no credential of that type; an existing credential of the type wins and the slot stays empty for the normal credential flow below. Add the tool with the credential slot omitted, then \`read_config\`.
-- Exception — when the user explicitly asks to run a tool on n8n credits, write \`{ "id": null, "name": "n8n credits", "__aiGatewayManaged": true }\` into that credential slot yourself: the server keeps it when the service is covered (even if the user has their own credential of the type) and removes it when not covered — check \`read_config\` after the write and resolve a real credential if it was removed.
-- The \`n8n credits\` managed credential IS the real, working credential — the tool executes through n8n's gateway on n8n credits, so NO separate API key is needed. It is NOT a placeholder and NOT "invalid for the service", even for a community node. For a slot \`read_config\` shows populated with it: the slot is fully connected and the tool WILL run. Do NOT call \`ask_credential\` for it; do NOT include it in \`finish_setup\`; NEVER clear, remove, or replace it via \`patch_config\`; and NEVER seek a "real" API key to swap in for it. Report the tool as ready, running on n8n credits — exactly like a managed model. Never tell the user the credential is "not connected"/"not set up" or that the tool "won't run until a credential is added".
-- Only for a required slot that \`read_config\` shows still empty after the write (a service n8n Connect does not cover) do you resolve a real credential: call \`ask_credential\` once before the config mutation for an addition to an existing agent. ${initial_build_prompt_1.INITIAL_BUILD_NOTE} After the trailing \`finish_setup\` resolves the credential, copy the returned credentials into \`node.credentials\` via \`patch_config\`; for resource-locator resolution follow \`agent-builder-resource-locators\` then. Pass the node's credential key as \`credentialSlot\`. On success, copy the returned \`credentials\` object directly to \`node.credentials\`. If skipped, still add the tool and omit only that credential slot.
+- Gateway credits cover many services, including some community nodes. Adding a node tool with its credential slot omitted triggers server-side assignment: for a covered service the server attaches the managed \`Gateway credits\` credential (\`{ id: null, name: "Gateway credits", __aiGatewayManaged: true }\`) to each required, eligible slot on write — but only when the project has no credential of that type; an existing credential of the type wins and the slot stays empty for the normal credential flow below. Add the tool with the credential slot omitted, then \`read_config\`.
+- Exception — when the user explicitly asks to run a tool on Gateway credits, write \`{ "id": null, "name": "Gateway credits", "__aiGatewayManaged": true }\` into that credential slot yourself: the server keeps it when the service is covered (even if the user has their own credential of the type) and removes it when not covered — check \`read_config\` after the write and resolve a real credential if it was removed.
+- The \`Gateway credits\` managed credential IS the real, working credential — the tool executes through n8n's gateway on Gateway credits, so NO separate API key is needed. It is NOT a placeholder and NOT "invalid for the service", even for a community node. For a slot \`read_config\` shows populated with it: the slot is fully connected and the tool WILL run. Do NOT call \`ask_credential\` for it; do NOT include it in \`finish_setup\`; NEVER clear, remove, or replace it via \`patch_config\`; and NEVER seek a "real" API key to swap in for it. Report the tool as ready, running on Gateway credits — exactly like a managed model. Never tell the user the credential is "not connected"/"not set up" or that the tool "won't run until a credential is added".
+- Only for a required slot that \`read_config\` shows still empty after the write (a service Gateway credits do not cover) do you resolve a real credential: call \`ask_credential\` once before the config mutation for an addition to an existing agent. ${initial_build_prompt_1.INITIAL_BUILD_NOTE} After the trailing \`finish_setup\` resolves the credential, copy the returned credentials into \`node.credentials\` via \`patch_config\`; for resource-locator resolution follow \`agent-builder-resource-locators\` then. Pass the node's credential key as \`credentialSlot\`. On success, copy the returned \`credentials\` object directly to \`node.credentials\`. If skipped, still add the tool and omit only that credential slot.
 - When the agent already has a chat channel configured and the tool needs the same
   credential type, \`ask_credential\` reuses the channel's credential automatically —
   do not ask the user to pick a different one.

package/dist/modules/agents/builder/skills/node-tools.skill.js.map

@@ -1 +1 @@
-{"version":3,"file":"node-tools.skill.js","sourceRoot":"","sources":["../../../../../src/modules/agents/builder/skills/node-tools.skill.ts"],"names":[],"mappings":";;;AACA,8CAAyD;AAEzD,0EAAqE;AAErE;IACC,OAAO;QACN,EAAE,EAAE,0BAA0B;QAC9B,IAAI,EAAE,0BAA0B;QAChC,WAAW,EACV,kOAAkO;QACnO,gBAAgB,EAAE;YACjB,qBAAqB;YACrB,cAAc;YACd,gBAAgB;YAChB,gBAAgB;YAChB,aAAa;YACb,cAAc;SACd;QACD,YAAY,EAAE;YACb,qBAAqB;YACrB,cAAc;YACd,gBAAgB;YAChB,gBAAgB;YAChB,8BAA8B;YAC9B,eAAe;YACf,aAAa;YACb,cAAc;YACd,cAAc;YACd,YAAY;SACZ;QACD,YAAY,EAAE;;;;;;;;;;;;;;;;;;;;;;;;gBAwBA,mCAAuB;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;2PA+BoN,yCAAkB;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;2DA2ClN;KACzD,CAAC;AACH,CAAC"}
+{"version":3,"file":"node-tools.skill.js","sourceRoot":"","sources":["../../../../../src/modules/agents/builder/skills/node-tools.skill.ts"],"names":[],"mappings":";;;AACA,8CAAyD;AAEzD,0EAAqE;AAErE;IACC,OAAO;QACN,EAAE,EAAE,0BAA0B;QAC9B,IAAI,EAAE,0BAA0B;QAChC,WAAW,EACV,kOAAkO;QACnO,gBAAgB,EAAE;YACjB,qBAAqB;YACrB,cAAc;YACd,gBAAgB;YAChB,gBAAgB;YAChB,aAAa;YACb,cAAc;SACd;QACD,YAAY,EAAE;YACb,qBAAqB;YACrB,cAAc;YACd,gBAAgB;YAChB,gBAAgB;YAChB,8BAA8B;YAC9B,eAAe;YACf,aAAa;YACb,cAAc;YACd,cAAc;YACd,YAAY;SACZ;QACD,YAAY,EAAE;;;;;;;;;;;;;;;;;;;;;;;;gBAwBA,mCAAuB;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;6PA+BsN,yCAAkB;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;2DA2CpN;KACzD,CAAC;AACH,CAAC"}

package/dist/modules/agents/builder/skills/sub-agents.skill.js

@@ -6,7 +6,7 @@ function subAgentsSkill() {
     return {
         id: 'agent-builder-sub-agents',
         name: 'Agent Builder Sub-Agents',
-        description: 'Use when configuring inline or saved sub-agent delegation for the target agent, selecting published same-project sub-agents, or changing subAgents.maxChildren.',
+        description: 'Use when configuring inline or saved sub-agent delegation for the target agent, selecting same-project sub-agents, or changing subAgents.maxChildren.',
         recommendedTools: ['list_sub_agents', 'ask_questions', 'read_config', 'patch_config'],
         allowedTools: [
             'list_sub_agents',
@@ -30,9 +30,8 @@ The target agent can call \`delegate_subagent\` with \`subAgentId: "inline"\`
 without any saved-agent refs. Inline subagents are ad-hoc child agents for
 one-off focused tasks.
 
-\`subAgents.agents\` is only for optional saved/published n8n Agent specialists
-that the target agent may select by id when they are a better fit than an inline
-subagent.
+\`subAgents.agents\` is only for optional saved n8n Agent specialists that the
+target agent may select by id when they are a better fit than an inline subagent.
 
 ## When to configure
 
@@ -43,19 +42,19 @@ subagent.
   delegated tasks in a run. It is not the same as \`config.toolCallConcurrency\`.
 - Do not create fields such as \`subAgents.maxConcurrentDelegations\`,
   \`delegationConcurrency\`, or \`delegateConcurrency\`.
-- Add saved subagent refs only when the user asks to use specific published
-  agents, reusable specialists, named helper agents, or saved-agent delegation.
+- Add saved subagent refs only when the user asks to use specific agents,
+  reusable specialists, named helper agents, or saved-agent delegation.
 
 ## Saved sub-agent workflow
 
-1. Call \`list_sub_agents\` to discover published same-project agents that can be
-   added. Do not write agent ids from memory, prose, or user-entered free text.
-2. If published agents are available and the user has not named exact agents,
+1. Call \`list_sub_agents\` to discover same-project agents that can be added.
+   Do not write agent ids from memory, prose, or user-entered free text.
+2. If agents are available and the user has not named exact agents,
    call \`ask_questions\` with one \`type: "multi"\` question whose \`options\`
    are the returned agent names. Map each selected option back to the
    matching \`agentId\` from the \`list_sub_agents\` result.
-3. If no published agents are available, do not configure saved subagents.
-   Inline delegation still works without saved-agent refs.
+3. If no agents are available, do not configure saved subagents. Inline
+   delegation still works without saved-agent refs.
 4. Determine the parent-owned routing guidance for each selected saved
    subagent. Store it as \`useWhen\`, for example
    \`{ "agentId": "<returned-agent-id>", "useWhen": "Use for billing-policy questions and invoice investigations." }\`.
@@ -80,6 +79,10 @@ Example patch flow:
 
 ## Rules
 
+- Saved subagents run their current draft in Preview/test runs and their
+  published version in production runs. Delegating to a never-published agent
+  fails in production, so tell the user to publish saved subagents the target
+  agent relies on before production use.
 - If the resumed values include text that is not one of the listed agent ids,
   do not persist it as an agent id; ask a follow-up.
 - \`useWhen\` is relationship-specific routing guidance owned by this parent
@@ -121,7 +124,7 @@ Example shape:
 ## Verify
 
 - Inline delegation still works even when \`subAgents.agents\` is absent.
-- Saved refs use only returned same-project published agent ids.
+- Saved refs use only returned same-project agent ids.
 - New or updated saved refs include concrete \`useWhen\` guidance.
 - Any \`subAgents.maxChildren\` value stays within ${api_types_1.SUB_AGENT_MAX_CHILDREN_MIN} to ${api_types_1.SUB_AGENT_MAX_CHILDREN_MAX}.`,
     };

package/dist/modules/agents/builder/skills/sub-agents.skill.js.map

@@ -1 +1 @@
-{"version":3,"file":"sub-agents.skill.js","sourceRoot":"","sources":["../../../../../src/modules/agents/builder/skills/sub-agents.skill.ts"],"names":[],"mappings":";;;AACA,8CAIwB;AAExB;IACC,OAAO;QACN,EAAE,EAAE,0BAA0B;QAC9B,IAAI,EAAE,0BAA0B;QAChC,WAAW,EACV,iKAAiK;QAClK,gBAAgB,EAAE,CAAC,iBAAiB,EAAE,eAAe,EAAE,aAAa,EAAE,cAAc,CAAC;QACrF,YAAY,EAAE;YACb,iBAAiB;YACjB,eAAe;YACf,aAAa;YACb,cAAc;YACd,cAAc;SACd;QACD,YAAY,EAAE;;;;;;;;;;;;;;;;;;;;;;;sDAuBsC,sCAA0B,OAAO,sCAA0B,gCAAgC,0CAA8B;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;qDAqF1H,sCAA0B,OAAO,sCAA0B,GAAG;KACjH,CAAC;AACH,CAAC"}
+{"version":3,"file":"sub-agents.skill.js","sourceRoot":"","sources":["../../../../../src/modules/agents/builder/skills/sub-agents.skill.ts"],"names":[],"mappings":";;;AACA,8CAIwB;AAExB;IACC,OAAO;QACN,EAAE,EAAE,0BAA0B;QAC9B,IAAI,EAAE,0BAA0B;QAChC,WAAW,EACV,uJAAuJ;QACxJ,gBAAgB,EAAE,CAAC,iBAAiB,EAAE,eAAe,EAAE,aAAa,EAAE,cAAc,CAAC;QACrF,YAAY,EAAE;YACb,iBAAiB;YACjB,eAAe;YACf,aAAa;YACb,cAAc;YACd,cAAc;SACd;QACD,YAAY,EAAE;;;;;;;;;;;;;;;;;;;;;;sDAsBsC,sCAA0B,OAAO,sCAA0B,gCAAgC,0CAA8B;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;qDAyF1H,sCAA0B,OAAO,sCAA0B,GAAG;KACjH,CAAC;AACH,CAAC"}

package/dist/modules/agents/builder/skills/target-tasks.skill.js

@@ -88,9 +88,13 @@ Initial Build rules in your system prompt. Never create a placeholder or
 - For each new or replacement objective, fill every template section with
   run-specific details. Do not duplicate Agent Instructions or Skill bodies;
   name a configured capability only when it helps route the work.
-- Make sure the agent already has every tool the steps need (an integration,
-  node/workflow tool, or web search). If something is missing, add it to the agent
-  config first — a task can only use tools the agent already has.
+- Make sure the agent already has every capability the steps need (an
+  integration, node/workflow tool, or web search). Configured chat integrations
+  automatically generate their context and action tools in scheduled runs too.
+  If the task sends through a configured platform and the integration lists the
+  needed action (for example \`send_dm\`), use that action; do not add a
+  same-platform node, MCP, or workflow tool. Add a tool only when the exact
+  operation is not supplied by an existing integration.
 - Translate each cadence into a valid 5-field cron expression (e.g. daily 09:00
   -> "0 9 * * *"; weekdays 08:30 -> "30 8 * * 1-5"; hourly -> "0 * * * *").
   Keep this cadence out of the objective; only a data lookback window belongs
@@ -125,8 +129,15 @@ Initial Build rules in your system prompt. Never create a placeholder or
 - To disable or remove a task, edit \`config.tasks\` with \`patch_config\` (set
   \`enabled: false\`, or drop the ref). Changes take effect on the next
   \`publish_agent\`.
-- \`create_tasks\` does NOT add tools — if a task needs a tool the agent lacks,
-  add it to the config yourself first.
+- \`create_tasks\` does not add config tools. This does not mean a scheduled
+  message needs a messaging node: configured integrations supply their generated
+  action tools at runtime even when the task has no inbound conversation.
+- A proactive or scheduled send through a connected chat platform is not a
+  separate backend integration. Use the channel's \`send_dm\` or
+  \`send_channel_message\` action when available.
+- Never claim that a task requires a same-platform messaging node solely because
+  it starts without an inbound conversation. If such a redundant tool already
+  exists and the integration covers its action, remove it.
 - Do not call \`create_tasks\` once per task when several are ready; batch them
   into one call so the whole set is stored in a single round trip.`,
     };

package/dist/modules/agents/builder/skills/target-tasks.skill.js.map

@@ -1 +1 @@
-{"version":3,"file":"target-tasks.skill.js","sourceRoot":"","sources":["../../../../../src/modules/agents/builder/skills/target-tasks.skill.ts"],"names":[],"mappings":";;;AAEA,wEAAiG;AAEjG;IACC,OAAO;QACN,EAAE,EAAE,4BAA4B;QAChC,IAAI,EAAE,4BAA4B;QAClC,WAAW,EACV,kMAAkM;QACnM,gBAAgB,EAAE;YACjB,cAAc;YACd,YAAY;YACZ,aAAa;YACb,eAAe;YACf,aAAa;YACb,cAAc;YACd,eAAe;SACf;QACD,YAAY,EAAE;YACb,cAAc;YACd,YAAY;YACZ,aAAa;YACb,eAAe;YACf,aAAa;YACb,cAAc;YACd,cAAc;YACd,gBAAgB;YAChB,cAAc;YACd,gBAAgB;YAChB,gBAAgB;YAChB,eAAe;SACf;QACD,YAAY,EAAE;;;;;;;;;;;;;;;;;;;;;;;;;EAyBd,oDAA0B;;EAE1B,iDAAuB;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;mEAuE0C;KACjE,CAAC;AACH,CAAC"}
+{"version":3,"file":"target-tasks.skill.js","sourceRoot":"","sources":["../../../../../src/modules/agents/builder/skills/target-tasks.skill.ts"],"names":[],"mappings":";;;AAEA,wEAAiG;AAEjG;IACC,OAAO;QACN,EAAE,EAAE,4BAA4B;QAChC,IAAI,EAAE,4BAA4B;QAClC,WAAW,EACV,kMAAkM;QACnM,gBAAgB,EAAE;YACjB,cAAc;YACd,YAAY;YACZ,aAAa;YACb,eAAe;YACf,aAAa;YACb,cAAc;YACd,eAAe;SACf;QACD,YAAY,EAAE;YACb,cAAc;YACd,YAAY;YACZ,aAAa;YACb,eAAe;YACf,aAAa;YACb,cAAc;YACd,cAAc;YACd,gBAAgB;YAChB,cAAc;YACd,gBAAgB;YAChB,gBAAgB;YAChB,eAAe;SACf;QACD,YAAY,EAAE;;;;;;;;;;;;;;;;;;;;;;;;;EAyBd,oDAA0B;;EAE1B,iDAAuB;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;mEAkF0C;KACjE,CAAC;AACH,CAAC"}

package/dist/modules/agents/entities/agent-channel-status.entity.d.ts

@@ -0,0 +1,16 @@
+import { WithTimestamps } from '@n8n/db';
+import { type Relation } from '@n8n/typeorm';
+import { Agent } from './agent.entity';
+export type AgentChannelStatusValue = 'connected' | 'error';
+export declare class AgentChannelStatus extends WithTimestamps {
+    agentId: string;
+    agent: Relation<Agent>;
+    integrationType: string;
+    credentialId: string;
+    hostId: string;
+    status: AgentChannelStatusValue;
+    errorMessage: string | null;
+    attempts: number;
+    backoffUntil: Date | null;
+    expiresAt: Date | null;
+}

package/dist/modules/agents/entities/agent-channel-status.entity.js

@@ -0,0 +1,99 @@
+"use strict";
+var __decorate = (this && this.__decorate) || function (decorators, target, key, desc) {
+    var c = arguments.length, r = c < 3 ? target : desc === null ? desc = Object.getOwnPropertyDescriptor(target, key) : desc, d;
+    if (typeof Reflect === "object" && typeof Reflect.decorate === "function") r = Reflect.decorate(decorators, target, key, desc);
+    else for (var i = decorators.length - 1; i >= 0; i--) if (d = decorators[i]) r = (c < 3 ? d(r) : c > 3 ? d(target, key, r) : d(target, key)) || r;
+    return c > 3 && r && Object.defineProperty(target, key, r), r;
+};
+var __metadata = (this && this.__metadata) || function (k, v) {
+    if (typeof Reflect === "object" && typeof Reflect.metadata === "function") return Reflect.metadata(k, v);
+};
+Object.defineProperty(exports, "__esModule", { value: true });
+exports.AgentChannelStatus = void 0;
+const db_1 = require("@n8n/db");
+const typeorm_1 = require("@n8n/typeorm");
+const agent_entity_1 = require("./agent.entity");
+let AgentChannelStatus = class AgentChannelStatus extends db_1.WithTimestamps {
+};
+exports.AgentChannelStatus = AgentChannelStatus;
+__decorate([
+    (0, typeorm_1.PrimaryColumn)({
+        type: 'varchar',
+        length: 36,
+        comment: 'Agent that owns this channel',
+    }),
+    __metadata("design:type", String)
+], AgentChannelStatus.prototype, "agentId", void 0);
+__decorate([
+    (0, typeorm_1.ManyToOne)(() => agent_entity_1.Agent, { onDelete: 'CASCADE' }),
+    (0, typeorm_1.JoinColumn)({ name: 'agentId' }),
+    __metadata("design:type", Object)
+], AgentChannelStatus.prototype, "agent", void 0);
+__decorate([
+    (0, typeorm_1.PrimaryColumn)({
+        type: 'varchar',
+        length: 64,
+        comment: 'Chat integration platform for this channel',
+    }),
+    __metadata("design:type", String)
+], AgentChannelStatus.prototype, "integrationType", void 0);
+__decorate([
+    (0, typeorm_1.PrimaryColumn)({
+        type: 'varchar',
+        length: 36,
+        comment: 'Credential connection that backs this channel; no FK so a failure is still recordable after the credential is deleted',
+    }),
+    __metadata("design:type", String)
+], AgentChannelStatus.prototype, "credentialId", void 0);
+__decorate([
+    (0, typeorm_1.Index)(),
+    (0, typeorm_1.PrimaryColumn)({
+        type: 'varchar',
+        length: 128,
+        comment: 'Process that observed this; the only writer of this row',
+    }),
+    __metadata("design:type", String)
+], AgentChannelStatus.prototype, "hostId", void 0);
+__decorate([
+    (0, typeorm_1.Column)({
+        type: 'varchar',
+        length: 16,
+        comment: 'What this process last observed: connected or error',
+    }),
+    __metadata("design:type", String)
+], AgentChannelStatus.prototype, "status", void 0);
+__decorate([
+    (0, typeorm_1.Column)({
+        type: 'text',
+        nullable: true,
+        comment: 'Why this process could not start the channel; null once it succeeds',
+    }),
+    __metadata("design:type", Object)
+], AgentChannelStatus.prototype, "errorMessage", void 0);
+__decorate([
+    (0, typeorm_1.Column)({
+        type: 'int',
+        default: 0,
+        comment: 'Consecutive failed startup attempts by this process, reset on success',
+    }),
+    __metadata("design:type", Number)
+], AgentChannelStatus.prototype, "attempts", void 0);
+__decorate([
+    (0, db_1.DateTimeColumn)({
+        nullable: true,
+        comment: 'Earliest this process should retry; null when there is nothing to retry',
+    }),
+    __metadata("design:type", Object)
+], AgentChannelStatus.prototype, "backoffUntil", void 0);
+__decorate([
+    (0, typeorm_1.Index)(),
+    (0, db_1.DateTimeColumn)({
+        nullable: true,
+        comment: 'When this row stops counting unless its owner refreshes it; null never expires',
+    }),
+    __metadata("design:type", Object)
+], AgentChannelStatus.prototype, "expiresAt", void 0);
+exports.AgentChannelStatus = AgentChannelStatus = __decorate([
+    (0, typeorm_1.Entity)({ name: 'agent_channel_status' })
+], AgentChannelStatus);
+//# sourceMappingURL=agent-channel-status.entity.js.map

package/dist/modules/agents/entities/agent-channel-status.entity.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"agent-channel-status.entity.js","sourceRoot":"","sources":["../../../../src/modules/agents/entities/agent-channel-status.entity.ts"],"names":[],"mappings":";;;;;;;;;;;;AAAA,gCAAyD;AACzD,0CAQsB;AAEtB,iDAAuC;AAwBhC,IAAM,kBAAkB,GAAxB,MAAM,kBAAmB,SAAQ,mBAAc;CAiFrD,CAAA;;AA3EA;IALC,IAAA,uBAAa,EAAC;QACd,IAAI,EAAE,SAAS;QACf,MAAM,EAAE,EAAE;QACV,OAAO,EAAE,8BAA8B;KACvC,CAAC;;mDACc;AAIhB;IAFC,IAAA,mBAAS,EAAC,GAAG,EAAE,CAAC,oBAAK,EAAE,EAAE,QAAQ,EAAE,SAAS,EAAE,CAAC;IAC/C,IAAA,oBAAU,EAAC,EAAE,IAAI,EAAE,SAAS,EAAE,CAAC;;iDACT;AAOvB;IALC,IAAA,uBAAa,EAAC;QACd,IAAI,EAAE,SAAS;QACf,MAAM,EAAE,EAAE;QACV,OAAO,EAAE,4CAA4C;KACrD,CAAC;;2DACsB;AAQxB;IANC,IAAA,uBAAa,EAAC;QACd,IAAI,EAAE,SAAS;QACf,MAAM,EAAE,EAAE;QACV,OAAO,EACN,uHAAuH;KACxH,CAAC;;wDACmB;AASrB;IANC,IAAA,eAAK,GAAE;IACP,IAAA,uBAAa,EAAC;QACd,IAAI,EAAE,SAAS;QACf,MAAM,EAAE,GAAG;QACX,OAAO,EAAE,yDAAyD;KAClE,CAAC;;kDACa;AAOf;IALC,IAAA,gBAAM,EAAC;QACP,IAAI,EAAE,SAAS;QACf,MAAM,EAAE,EAAE;QACV,OAAO,EAAE,qDAAqD;KAC9D,CAAC;;kDAC8B;AAOhC;IALC,IAAA,gBAAM,EAAC;QACP,IAAI,EAAE,MAAM;QACZ,QAAQ,EAAE,IAAI;QACd,OAAO,EAAE,qEAAqE;KAC9E,CAAC;;wDAC0B;AAO5B;IALC,IAAA,gBAAM,EAAC;QACP,IAAI,EAAE,KAAK;QACX,OAAO,EAAE,CAAC;QACV,OAAO,EAAE,uEAAuE;KAChF,CAAC;;oDACe;AAUjB;IAJC,IAAA,mBAAc,EAAC;QACf,QAAQ,EAAE,IAAI;QACd,OAAO,EAAE,yEAAyE;KAClF,CAAC;;wDACwB;AAe1B;IALC,IAAA,eAAK,GAAE;IACP,IAAA,mBAAc,EAAC;QACf,QAAQ,EAAE,IAAI;QACd,OAAO,EAAE,gFAAgF;KACzF,CAAC;;qDACqB;6BAhFX,kBAAkB;IAD9B,IAAA,gBAAM,EAAC,EAAE,IAAI,EAAE,sBAAsB,EAAE,CAAC;GAC5B,kBAAkB,CAiF9B"}

package/dist/modules/agents/instance-ai-builder-delegate.adapter.d.ts

@@ -1,13 +1,13 @@
 import type { CredentialProvider } from '@n8n/agents';
 import type { User } from '@n8n/db';
-import type { InstanceAiBuilderDelegate } from '@n8n/instance-ai';
+import { type InstanceAiBuilderDelegate, type InstanceAiCredentialService } from '@n8n/instance-ai';
 import { AgentConfigService } from './agent-config.service';
 import { AgentSkillsService } from './agent-skills.service';
 import { AgentsService } from './agents.service';
 import { AgentsBuilderService } from './builder/agents-builder.service';
 import { N8nMemory } from './integrations/n8n-memory';
 import { AgentThreadRepository } from './repositories/agent-thread.repository';
-export declare const INSTANCE_AI_BUILDER_ADDENDUM = "## Instance AI session rules\n\nYou are running as a sub-agent inside n8n's instance AI chat; the user sees your questions as chat cards.\n\nPreview links work in this chat. Include a markdown Preview link after a successful build and when `call_agent` reports an unsupported interaction as `approval_required`, using the exact relative path from \"When To Build vs When To Converse\" (form: `[Preview](<path>)`). Do not invent absolute URLs. Do not omit the link and describe the path in plain text instead.\n\nYou can publish and unpublish the target agent with `publish_agent` and `unpublish_agent`. Never tell the user to open the agent editor and click Publish.\n\nThe Instance AI orchestrator can create workflows and data tables \u2014 never ask the user to create them manually. State missing prerequisites in your reply; the orchestrator will provision them and call you again.";
+export declare const INSTANCE_AI_BUILDER_ADDENDUM = "## Instance AI session rules\n\nYou are running as a sub-agent inside n8n's instance AI chat; the user sees your questions as chat cards.\n\nPreview links work in this chat. Include a markdown Preview link after a successful build and when `call_agent` reports an unsupported interaction as `approval_required`, using the exact relative path from \"When To Build vs When To Converse\" (form: `[Preview](<path>)`). Do not invent absolute URLs. Do not omit the link and describe the path in plain text instead.\n\nYou can publish and unpublish the target agent with `publish_agent` and `unpublish_agent`. Never tell the user to open the agent editor and click Publish.\n\nThe Instance AI orchestrator can create workflows and data tables \u2014 never ask the user to create them manually. For each missing artifact, call `report_required_artifact` with its concrete requirements before your final reply; the orchestrator will provision them and call you again when the Agent needs the result.\n\nSome requested chat platforms do not have a native Agent integration. In that case, finish the Agent without adding same-platform messaging nodes as Agent tools, then report an `agent-entrypoint` workflow. It must use the platform trigger, pass the incoming message and a stable conversation identifier into Message an Agent with a custom session key, then send the Agent's text response back through the platform. This workflow invokes the Agent and must never be attached to the Agent as a workflow tool.";
 export declare class InstanceAiBuilderDelegateAdapterService {
     private readonly agentsService;
     private readonly agentsBuilderService;
@@ -17,6 +17,6 @@ export declare class InstanceAiBuilderDelegateAdapterService {
     private readonly agentSkills;
     constructor(agentsService: AgentsService, agentsBuilderService: AgentsBuilderService, n8nMemory: N8nMemory, agentThreadRepository: AgentThreadRepository, agentConfig: AgentConfigService, agentSkills: AgentSkillsService);
     private buildSubAgentSession;
-    createDelegate(user: User, projectId: string, credentialProvider: CredentialProvider): InstanceAiBuilderDelegate;
+    createDelegate(user: User, projectId: string, credentialProvider: CredentialProvider, credentialService: InstanceAiCredentialService): InstanceAiBuilderDelegate;
     deleteBuilderSessions(instanceAiThreadId: string): Promise<void>;
 }

package/dist/modules/agents/instance-ai-builder-delegate.adapter.js

@@ -31,15 +31,21 @@ Preview links work in this chat. Include a markdown Preview link after a success
 
 You can publish and unpublish the target agent with \`publish_agent\` and \`unpublish_agent\`. Never tell the user to open the agent editor and click Publish.
 
-The Instance AI orchestrator can create workflows and data tables — never ask the user to create them manually. State missing prerequisites in your reply; the orchestrator will provision them and call you again.`;
+The Instance AI orchestrator can create workflows and data tables — never ask the user to create them manually. For each missing artifact, call \`report_required_artifact\` with its concrete requirements before your final reply; the orchestrator will provision them and call you again when the Agent needs the result.
+
+Some requested chat platforms do not have a native Agent integration. In that case, finish the Agent without adding same-platform messaging nodes as Agent tools, then report an \`agent-entrypoint\` workflow. It must use the platform trigger, pass the incoming message and a stable conversation identifier into Message an Agent with a custom session key, then send the Agent's text response back through the platform. This workflow invokes the Agent and must never be attached to the Agent as a workflow tool.`;
 function isTextDeltaChunk(chunk) {
     return chunk.type === 'text-delta';
 }
-function toBuilderTurnStream(chunks) {
+function toBuilderTurnStream(chunks, requiredArtifacts) {
     let resolveText;
     const text = new Promise((resolve) => {
         resolveText = resolve;
     });
+    let resolveRequiredArtifacts;
+    const requiredArtifactsPromise = new Promise((resolve) => {
+        resolveRequiredArtifacts = resolve;
+    });
     let acc = '';
     async function* pump() {
         try {
@@ -51,9 +57,10 @@ function toBuilderTurnStream(chunks) {
         }
         finally {
             resolveText(acc);
+            resolveRequiredArtifacts([...requiredArtifacts]);
         }
     }
-    return { fullStream: pump(), text };
+    return { fullStream: pump(), text, requiredArtifacts: requiredArtifactsPromise };
 }
 let InstanceAiBuilderDelegateAdapterService = class InstanceAiBuilderDelegateAdapterService {
     constructor(agentsService, agentsBuilderService, n8nMemory, agentThreadRepository, agentConfig, agentSkills) {
@@ -64,7 +71,7 @@ let InstanceAiBuilderDelegateAdapterService = class InstanceAiBuilderDelegateAda
         this.agentConfig = agentConfig;
         this.agentSkills = agentSkills;
     }
-    buildSubAgentSession(session) {
+    buildSubAgentSession(session, onRequiredArtifact) {
         return {
             threadId: session.threadId,
             hostThreadId: session.hostThreadId,
@@ -75,9 +82,10 @@ let InstanceAiBuilderDelegateAdapterService = class InstanceAiBuilderDelegateAda
             ...(session.memoryTaskObserver ? { memoryTaskObserver: session.memoryTaskObserver } : {}),
             abortSignal: session.abortSignal,
             ...(session.mcpTools ? { mcpTools: session.mcpTools } : {}),
+            onRequiredArtifact,
         };
     }
-    createDelegate(user, projectId, credentialProvider) {
+    createDelegate(user, projectId, credentialProvider, credentialService) {
         const assertProjectScope = async (scope) => {
             if (!(await (0, check_access_1.userHasScopes)(user, [scope], false, { projectId }))) {
                 throw new forbidden_error_1.ForbiddenError('You do not have permission to access agents in this project.');
@@ -94,11 +102,13 @@ let InstanceAiBuilderDelegateAdapterService = class InstanceAiBuilderDelegateAda
             },
             streamBuild: async (agentId, message, session) => {
                 await assertProjectScope('agent:update');
-                return toBuilderTurnStream(this.agentsBuilderService.buildAgent(agentId, projectId, message, credentialProvider, user, this.buildSubAgentSession(session)));
+                const requiredArtifacts = [];
+                return toBuilderTurnStream(this.agentsBuilderService.buildAgent(agentId, projectId, message, credentialProvider, credentialService, user, this.buildSubAgentSession(session, (artifact) => requiredArtifacts.push(artifact))), requiredArtifacts);
             },
             resumeBuild: async (agentId, resume, session) => {
                 await assertProjectScope('agent:update');
-                return toBuilderTurnStream(this.agentsBuilderService.resumeBuild(agentId, projectId, resume.runId, resume.toolCallId, resume.resumeData, credentialProvider, user, this.buildSubAgentSession(session)));
+                const requiredArtifacts = [];
+                return toBuilderTurnStream(this.agentsBuilderService.resumeBuild(agentId, projectId, resume.runId, resume.toolCallId, resume.resumeData, credentialProvider, credentialService, user, this.buildSubAgentSession(session, (artifact) => requiredArtifacts.push(artifact))), requiredArtifacts);
             },
             findOpenSuspensions: async (agentId, session) => {
                 await assertProjectScope('agent:update');

package/dist/modules/agents/instance-ai-builder-delegate.adapter.js.map

@@ -1 +1 @@
-{"version":3,"file":"instance-ai-builder-delegate.adapter.js","sourceRoot":"","sources":["../../../src/modules/agents/instance-ai-builder-delegate.adapter.ts"],"names":[],"mappings":";;;;;;;;;;;;AAEA,gCAAkC;AAClC,kDAAiE;AAOjE,0CAAoC;AACpC,+CAAyC;AAEzC,8EAA0E;AAC1E,gEAA8D;AAE9D,iEAA4D;AAC5D,iEAA4D;AAC5D,qDAAiD;AACjD,6EAAwE;AAExE,0DAAsD;AACtD,oFAA+E;AAC/E,iEAA+D;AAGlD,QAAA,4BAA4B,GAAG;;;;;;;;oNAQwK,CAAC;AAErN,SAAS,gBAAgB,CACxB,KAAkB;IAElB,OAAO,KAAK,CAAC,IAAI,KAAK,YAAY,CAAC;AACpC,CAAC;AAKD,SAAS,mBAAmB,CAAC,MAAmC;IAC/D,IAAI,WAAmC,CAAC;IACxC,MAAM,IAAI,GAAG,IAAI,OAAO,CAAS,CAAC,OAAO,EAAE,EAAE;QAC5C,WAAW,GAAG,OAAO,CAAC;IACvB,CAAC,CAAC,CAAC;IACH,IAAI,GAAG,GAAG,EAAE,CAAC;IAEb,KAAK,SAAS,CAAC,CAAC,IAAI;QACnB,IAAI,CAAC;YACJ,IAAI,KAAK,EAAE,MAAM,KAAK,IAAI,MAAM,EAAE,CAAC;gBAClC,IAAI,gBAAgB,CAAC,KAAK,CAAC;oBAAE,GAAG,IAAI,KAAK,CAAC,KAAK,CAAC;gBAChD,MAAM,KAAK,CAAC;YACb,CAAC;QACF,CAAC;gBAAS,CAAC;YACV,WAAW,CAAC,GAAG,CAAC,CAAC;QAClB,CAAC;IACF,CAAC;IAED,OAAO,EAAE,UAAU,EAAE,IAAI,EAAE,EAAE,IAAI,EAAE,CAAC;AACrC,CAAC;AAcM,IAAM,uCAAuC,GAA7C,MAAM,uCAAuC;IACnD,YACkB,aAA4B,EAC5B,oBAA0C,EAC1C,SAAoB,EACpB,qBAA4C,EAC5C,WAA+B,EAC/B,WAA+B;6BAL/B,aAAa;oCACb,oBAAoB;yBACpB,SAAS;qCACT,qBAAqB;2BACrB,WAAW;2BACX,WAAW;IAC1B,CAAC;IAGI,oBAAoB,CAAC,OAA+B;QAC3D,OAAO;YACN,QAAQ,EAAE,OAAO,CAAC,QAAQ;YAC1B,YAAY,EAAE,OAAO,CAAC,YAAY;YAClC,KAAK,EAAE,OAAO,CAAC,KAAK;YACpB,oBAAoB,EAAE,QAAA,4BAA4B;YAClD,WAAW,EAAE,OAAO,CAAC,WAAW;YAChC,GAAG,CAAC,OAAO,CAAC,SAAS,CAAC,CAAC,CAAC,EAAE,SAAS,EAAE,OAAO,CAAC,SAAS,EAAE,CAAC,CAAC,CAAC,EAAE,CAAC;YAC9D,GAAG,CAAC,OAAO,CAAC,kBAAkB,CAAC,CAAC,CAAC,EAAE,kBAAkB,EAAE,OAAO,CAAC,kBAAkB,EAAE,CAAC,CAAC,CAAC,EAAE,CAAC;YACzF,WAAW,EAAE,OAAO,CAAC,WAAW;YAChC,GAAG,CAAC,OAAO,CAAC,QAAQ,CAAC,CAAC,CAAC,EAAE,QAAQ,EAAE,OAAO,CAAC,QAAQ,EAAE,CAAC,CAAC,CAAC,EAAE,CAAC;SAC3D,CAAC;IACH,CAAC;IAED,cAAc,CACb,IAAU,EACV,SAAiB,EACjB,kBAAsC;QAMtC,MAAM,kBAAkB,GAAG,KAAK,EAAE,KAAY,EAAiB,EAAE;YAChE,IAAI,CAAC,CAAC,MAAM,IAAA,4BAAa,EAAC,IAAI,EAAE,CAAC,KAAK,CAAC,EAAE,KAAK,EAAE,EAAE,SAAS,EAAE,CAAC,CAAC,EAAE,CAAC;gBACjE,MAAM,IAAI,gCAAc,CAAC,8DAA8D,CAAC,CAAC;YAC1F,CAAC;QACF,CAAC,CAAC;QAEF,OAAO;YACN,WAAW,EAAE,KAAK,EAAE,IAAI,EAAE,EAAE,EAAE,EAAE;gBAC/B,MAAM,kBAAkB,CAAC,cAAc,CAAC,CAAC;gBACzC,MAAM,KAAK,GAAG,MAAM,IAAI,CAAC,aAAa,CAAC,MAAM,CAAC,SAAS,EAAE,IAAI,EAAE;oBAC9D,EAAE;oBACF,4BAA4B,EAAE,IAAI;iBAClC,CAAC,CAAC;gBACH,OAAO,EAAE,OAAO,EAAE,KAAK,CAAC,EAAE,EAAE,SAAS,EAAE,CAAC;YACzC,CAAC;YAED,WAAW,EAAE,KAAK,EAAE,OAAO,EAAE,OAAO,EAAE,OAAO,EAAE,EAAE;gBAChD,MAAM,kBAAkB,CAAC,cAAc,CAAC,CAAC;gBACzC,OAAO,mBAAmB,CACzB,IAAI,CAAC,oBAAoB,CAAC,UAAU,CACnC,OAAO,EACP,SAAS,EACT,OAAO,EACP,kBAAkB,EAClB,IAAI,EACJ,IAAI,CAAC,oBAAoB,CAAC,OAAO,CAAC,CAClC,CACD,CAAC;YACH,CAAC;YAED,WAAW,EAAE,KAAK,EAAE,OAAO,EAAE,MAAM,EAAE,OAAO,EAAE,EAAE;gBAC/C,MAAM,kBAAkB,CAAC,cAAc,CAAC,CAAC;gBACzC,OAAO,mBAAmB,CACzB,IAAI,CAAC,oBAAoB,CAAC,WAAW,CACpC,OAAO,EACP,SAAS,EACT,MAAM,CAAC,KAAK,EACZ,MAAM,CAAC,UAAU,EACjB,MAAM,CAAC,UAAU,EACjB,kBAAkB,EAClB,IAAI,EACJ,IAAI,CAAC,oBAAoB,CAAC,OAAO,CAAC,CAClC,CACD,CAAC;YACH,CAAC;YAED,mBAAmB,EAAE,KAAK,EAAE,OAAO,EAAE,OAAO,EAAE,EAAE;gBAC/C,MAAM,kBAAkB,CAAC,cAAc,CAAC,CAAC;gBACzC,MAAM,UAAU,GAAG,MAAM,IAAI,CAAC,oBAAoB,CAAC,2BAA2B,CAC7E,OAAO,EACP,OAAO,CAAC,QAAQ,CAChB,CAAC;gBACF,IAAI,CAAC,UAAU;oBAAE,OAAO,EAAE,CAAC;gBAC3B,OAAO,MAAM,CAAC,MAAM,CAAC,UAAU,CAAC,gBAAgB,IAAI,EAAE,CAAC;qBACrD,MAAM,CAAC,CAAC,EAAE,EAAE,EAAE,CAAC,EAAE,CAAC,SAAS,CAAC;qBAC5B,GAAG,CAAC,CAAC,EAAE,EAAE,EAAE,CAAC,CAAC,EAAE,KAAK,EAAE,EAAE,CAAC,KAAK,EAAE,UAAU,EAAE,EAAE,CAAC,UAAU,EAAE,CAAC,CAAC,CAAC;YACjE,CAAC;YAED,oBAAoB,EAAE,KAAK,EAAE,OAAO,EAAE,KAAK,EAAE,EAAE;gBAC9C,MAAM,kBAAkB,CAAC,cAAc,CAAC,CAAC;gBACzC,MAAM,IAAI,CAAC,oBAAoB,CAAC,gBAAgB,CAAC,OAAO,EAAE,KAAK,CAAC,CAAC;YAClE,CAAC;YAED,UAAU,EAAE,KAAK,IAAI,EAAE;gBACtB,MAAM,kBAAkB,CAAC,YAAY,CAAC,CAAC;gBACvC,MAAM,MAAM,GAAG,MAAM,IAAI,CAAC,aAAa,CAAC,eAAe,CAAC,SAAS,CAAC,CAAC;gBACnE,OAAO,MAAM,CAAC,GAAG,CAAC,CAAC,KAAK,EAAE,EAAE,CAAC,CAAC;oBAC7B,OAAO,EAAE,KAAK,CAAC,EAAE;oBACjB,IAAI,EAAE,KAAK,CAAC,IAAI;oBAChB,SAAS,EAAE,KAAK,CAAC,eAAe,KAAK,IAAI;oBACzC,SAAS,EAAE,KAAK,CAAC,SAAS,CAAC,WAAW,EAAE;iBACxC,CAAC,CAAC,CAAC;YACL,CAAC;YAED,gBAAgB,EAAE,KAAK,EAAE,OAAO,EAAE,EAAE;gBACnC,MAAM,kBAAkB,CAAC,YAAY,CAAC,CAAC;gBACvC,OAAO,CAAC,MAAM,IAAI,CAAC,aAAa,CAAC,QAAQ,CAAC,OAAO,EAAE,SAAS,CAAC,CAAC,EAAE,IAAI,CAAC;YACtE,CAAC;YACD,iBAAiB,EAAE,KAAK,EAAE,OAAO,EAAE,EAAE;gBACpC,MAAM,kBAAkB,CAAC,YAAY,CAAC,CAAC;gBAIvC,MAAM,MAAM,GAAG,MAAM,IAAI,CAAC,WAAW,CAAC,SAAS,CAAC,OAAO,EAAE,SAAS,CAAC,CAAC,KAAK,CAAC,CAAC,KAAK,EAAE,EAAE;oBACnF,IAAI,KAAK,YAAY,wBAAS;wBAAE,OAAO,IAAI,CAAC;oBAC5C,MAAM,KAAK,CAAC;gBACb,CAAC,CAAC,CAAC;gBACH,IAAI,CAAC,MAAM;oBAAE,OAAO,IAAI,CAAC;gBACzB,OAAO;oBACN,MAAM;oBACN,MAAM,EAAE,MAAM,IAAI,CAAC,WAAW,CAAC,UAAU,CAAC,OAAO,EAAE,SAAS,CAAC;oBAE7D,UAAU,EAAE,IAAA,sCAAkB,EAAC,MAAM,CAAC;iBACtC,CAAC;YACH,CAAC;SACD,CAAC;IACH,CAAC;IAUD,KAAK,CAAC,qBAAqB,CAAC,kBAA0B;QACrD,MAAM,MAAM,GAAG,IAAA,2CAA6B,EAAC,kBAAkB,CAAC,CAAC;QACjE,MAAM,OAAO,GAAG,MAAM,IAAI,CAAC,qBAAqB,CAAC,IAAI,CAAC;YACrD,MAAM,EAAE,EAAE,EAAE,EAAE,IAAI,EAAE;YACpB,KAAK,EAAE,EAAE,EAAE,EAAE,IAAA,cAAI,EAAC,GAAG,MAAM,GAAG,CAAC,EAAE;SACjC,CAAC,CAAC;QACH,KAAK,MAAM,EAAE,EAAE,EAAE,IAAI,OAAO,EAAE,CAAC;YAE9B,MAAM,MAAM,GAAG,IAAI,CAAC,SAAS,CAAC,iBAAiB,CAAC,EAAE,CAAC,KAAK,CAAC,MAAM,CAAC,MAAM,CAAC,CAAC,CAAC;YACzE,MAAM,MAAM,CAAC,sBAAsB,CAAC,EAAE,CAAC,CAAC;YACxC,MAAM,MAAM,CAAC,YAAY,CAAC,EAAE,CAAC,CAAC;QAC/B,CAAC;IACF,CAAC;CACD,CAAA;;kDAzJY,uCAAuC;IADnD,IAAA,YAAO,GAAE;qCAGwB,8BAAa,EACN,6CAAoB,EAC/B,sBAAS,EACG,+CAAqB,EAC/B,yCAAkB,EAClB,yCAAkB;GAPrC,uCAAuC,CAyJnD"}
+{"version":3,"file":"instance-ai-builder-delegate.adapter.js","sourceRoot":"","sources":["../../../src/modules/agents/instance-ai-builder-delegate.adapter.ts"],"names":[],"mappings":";;;;;;;;;;;;AAEA,gCAAkC;AAClC,kDAO0B;AAE1B,0CAAoC;AACpC,+CAAyC;AAEzC,8EAA0E;AAC1E,gEAA8D;AAE9D,iEAA4D;AAC5D,iEAA4D;AAC5D,qDAAiD;AACjD,6EAAwE;AAExE,0DAAsD;AACtD,oFAA+E;AAC/E,iEAA+D;AAGlD,QAAA,4BAA4B,GAAG;;;;;;;;;;6fAUid,CAAC;AAE9f,SAAS,gBAAgB,CACxB,KAAkB;IAElB,OAAO,KAAK,CAAC,IAAI,KAAK,YAAY,CAAC;AACpC,CAAC;AAKD,SAAS,mBAAmB,CAC3B,MAAmC,EACnC,iBAA4C;IAE5C,IAAI,WAAmC,CAAC;IACxC,MAAM,IAAI,GAAG,IAAI,OAAO,CAAS,CAAC,OAAO,EAAE,EAAE;QAC5C,WAAW,GAAG,OAAO,CAAC;IACvB,CAAC,CAAC,CAAC;IACH,IAAI,wBAAwE,CAAC;IAC7E,MAAM,wBAAwB,GAAG,IAAI,OAAO,CAA4B,CAAC,OAAO,EAAE,EAAE;QACnF,wBAAwB,GAAG,OAAO,CAAC;IACpC,CAAC,CAAC,CAAC;IACH,IAAI,GAAG,GAAG,EAAE,CAAC;IAEb,KAAK,SAAS,CAAC,CAAC,IAAI;QACnB,IAAI,CAAC;YACJ,IAAI,KAAK,EAAE,MAAM,KAAK,IAAI,MAAM,EAAE,CAAC;gBAClC,IAAI,gBAAgB,CAAC,KAAK,CAAC;oBAAE,GAAG,IAAI,KAAK,CAAC,KAAK,CAAC;gBAChD,MAAM,KAAK,CAAC;YACb,CAAC;QACF,CAAC;gBAAS,CAAC;YACV,WAAW,CAAC,GAAG,CAAC,CAAC;YACjB,wBAAwB,CAAC,CAAC,GAAG,iBAAiB,CAAC,CAAC,CAAC;QAClD,CAAC;IACF,CAAC;IAED,OAAO,EAAE,UAAU,EAAE,IAAI,EAAE,EAAE,IAAI,EAAE,iBAAiB,EAAE,wBAAwB,EAAE,CAAC;AAClF,CAAC;AAcM,IAAM,uCAAuC,GAA7C,MAAM,uCAAuC;IACnD,YACkB,aAA4B,EAC5B,oBAA0C,EAC1C,SAAoB,EACpB,qBAA4C,EAC5C,WAA+B,EAC/B,WAA+B;6BAL/B,aAAa;oCACb,oBAAoB;yBACpB,SAAS;qCACT,qBAAqB;2BACrB,WAAW;2BACX,WAAW;IAC1B,CAAC;IAGI,oBAAoB,CAC3B,OAA+B,EAC/B,kBAA+D;QAE/D,OAAO;YACN,QAAQ,EAAE,OAAO,CAAC,QAAQ;YAC1B,YAAY,EAAE,OAAO,CAAC,YAAY;YAClC,KAAK,EAAE,OAAO,CAAC,KAAK;YACpB,oBAAoB,EAAE,QAAA,4BAA4B;YAClD,WAAW,EAAE,OAAO,CAAC,WAAW;YAChC,GAAG,CAAC,OAAO,CAAC,SAAS,CAAC,CAAC,CAAC,EAAE,SAAS,EAAE,OAAO,CAAC,SAAS,EAAE,CAAC,CAAC,CAAC,EAAE,CAAC;YAC9D,GAAG,CAAC,OAAO,CAAC,kBAAkB,CAAC,CAAC,CAAC,EAAE,kBAAkB,EAAE,OAAO,CAAC,kBAAkB,EAAE,CAAC,CAAC,CAAC,EAAE,CAAC;YACzF,WAAW,EAAE,OAAO,CAAC,WAAW;YAChC,GAAG,CAAC,OAAO,CAAC,QAAQ,CAAC,CAAC,CAAC,EAAE,QAAQ,EAAE,OAAO,CAAC,QAAQ,EAAE,CAAC,CAAC,CAAC,EAAE,CAAC;YAC3D,kBAAkB;SAClB,CAAC;IACH,CAAC;IAED,cAAc,CACb,IAAU,EACV,SAAiB,EACjB,kBAAsC,EACtC,iBAA8C;QAM9C,MAAM,kBAAkB,GAAG,KAAK,EAAE,KAAY,EAAiB,EAAE;YAChE,IAAI,CAAC,CAAC,MAAM,IAAA,4BAAa,EAAC,IAAI,EAAE,CAAC,KAAK,CAAC,EAAE,KAAK,EAAE,EAAE,SAAS,EAAE,CAAC,CAAC,EAAE,CAAC;gBACjE,MAAM,IAAI,gCAAc,CAAC,8DAA8D,CAAC,CAAC;YAC1F,CAAC;QACF,CAAC,CAAC;QAEF,OAAO;YACN,WAAW,EAAE,KAAK,EAAE,IAAI,EAAE,EAAE,EAAE,EAAE;gBAC/B,MAAM,kBAAkB,CAAC,cAAc,CAAC,CAAC;gBACzC,MAAM,KAAK,GAAG,MAAM,IAAI,CAAC,aAAa,CAAC,MAAM,CAAC,SAAS,EAAE,IAAI,EAAE;oBAC9D,EAAE;oBACF,4BAA4B,EAAE,IAAI;iBAClC,CAAC,CAAC;gBACH,OAAO,EAAE,OAAO,EAAE,KAAK,CAAC,EAAE,EAAE,SAAS,EAAE,CAAC;YACzC,CAAC;YAED,WAAW,EAAE,KAAK,EAAE,OAAO,EAAE,OAAO,EAAE,OAAO,EAAE,EAAE;gBAChD,MAAM,kBAAkB,CAAC,cAAc,CAAC,CAAC;gBACzC,MAAM,iBAAiB,GAA8B,EAAE,CAAC;gBACxD,OAAO,mBAAmB,CACzB,IAAI,CAAC,oBAAoB,CAAC,UAAU,CACnC,OAAO,EACP,SAAS,EACT,OAAO,EACP,kBAAkB,EAClB,iBAAiB,EACjB,IAAI,EACJ,IAAI,CAAC,oBAAoB,CAAC,OAAO,EAAE,CAAC,QAAQ,EAAE,EAAE,CAAC,iBAAiB,CAAC,IAAI,CAAC,QAAQ,CAAC,CAAC,CAClF,EACD,iBAAiB,CACjB,CAAC;YACH,CAAC;YAED,WAAW,EAAE,KAAK,EAAE,OAAO,EAAE,MAAM,EAAE,OAAO,EAAE,EAAE;gBAC/C,MAAM,kBAAkB,CAAC,cAAc,CAAC,CAAC;gBACzC,MAAM,iBAAiB,GAA8B,EAAE,CAAC;gBACxD,OAAO,mBAAmB,CACzB,IAAI,CAAC,oBAAoB,CAAC,WAAW,CACpC,OAAO,EACP,SAAS,EACT,MAAM,CAAC,KAAK,EACZ,MAAM,CAAC,UAAU,EACjB,MAAM,CAAC,UAAU,EACjB,kBAAkB,EAClB,iBAAiB,EACjB,IAAI,EACJ,IAAI,CAAC,oBAAoB,CAAC,OAAO,EAAE,CAAC,QAAQ,EAAE,EAAE,CAAC,iBAAiB,CAAC,IAAI,CAAC,QAAQ,CAAC,CAAC,CAClF,EACD,iBAAiB,CACjB,CAAC;YACH,CAAC;YAED,mBAAmB,EAAE,KAAK,EAAE,OAAO,EAAE,OAAO,EAAE,EAAE;gBAC/C,MAAM,kBAAkB,CAAC,cAAc,CAAC,CAAC;gBACzC,MAAM,UAAU,GAAG,MAAM,IAAI,CAAC,oBAAoB,CAAC,2BAA2B,CAC7E,OAAO,EACP,OAAO,CAAC,QAAQ,CAChB,CAAC;gBACF,IAAI,CAAC,UAAU;oBAAE,OAAO,EAAE,CAAC;gBAC3B,OAAO,MAAM,CAAC,MAAM,CAAC,UAAU,CAAC,gBAAgB,IAAI,EAAE,CAAC;qBACrD,MAAM,CAAC,CAAC,EAAE,EAAE,EAAE,CAAC,EAAE,CAAC,SAAS,CAAC;qBAC5B,GAAG,CAAC,CAAC,EAAE,EAAE,EAAE,CAAC,CAAC,EAAE,KAAK,EAAE,EAAE,CAAC,KAAK,EAAE,UAAU,EAAE,EAAE,CAAC,UAAU,EAAE,CAAC,CAAC,CAAC;YACjE,CAAC;YAED,oBAAoB,EAAE,KAAK,EAAE,OAAO,EAAE,KAAK,EAAE,EAAE;gBAC9C,MAAM,kBAAkB,CAAC,cAAc,CAAC,CAAC;gBACzC,MAAM,IAAI,CAAC,oBAAoB,CAAC,gBAAgB,CAAC,OAAO,EAAE,KAAK,CAAC,CAAC;YAClE,CAAC;YAED,UAAU,EAAE,KAAK,IAAI,EAAE;gBACtB,MAAM,kBAAkB,CAAC,YAAY,CAAC,CAAC;gBACvC,MAAM,MAAM,GAAG,MAAM,IAAI,CAAC,aAAa,CAAC,eAAe,CAAC,SAAS,CAAC,CAAC;gBACnE,OAAO,MAAM,CAAC,GAAG,CAAC,CAAC,KAAK,EAAE,EAAE,CAAC,CAAC;oBAC7B,OAAO,EAAE,KAAK,CAAC,EAAE;oBACjB,IAAI,EAAE,KAAK,CAAC,IAAI;oBAChB,SAAS,EAAE,KAAK,CAAC,eAAe,KAAK,IAAI;oBACzC,SAAS,EAAE,KAAK,CAAC,SAAS,CAAC,WAAW,EAAE;iBACxC,CAAC,CAAC,CAAC;YACL,CAAC;YAED,gBAAgB,EAAE,KAAK,EAAE,OAAO,EAAE,EAAE;gBACnC,MAAM,kBAAkB,CAAC,YAAY,CAAC,CAAC;gBACvC,OAAO,CAAC,MAAM,IAAI,CAAC,aAAa,CAAC,QAAQ,CAAC,OAAO,EAAE,SAAS,CAAC,CAAC,EAAE,IAAI,CAAC;YACtE,CAAC;YACD,iBAAiB,EAAE,KAAK,EAAE,OAAO,EAAE,EAAE;gBACpC,MAAM,kBAAkB,CAAC,YAAY,CAAC,CAAC;gBAIvC,MAAM,MAAM,GAAG,MAAM,IAAI,CAAC,WAAW,CAAC,SAAS,CAAC,OAAO,EAAE,SAAS,CAAC,CAAC,KAAK,CAAC,CAAC,KAAK,EAAE,EAAE;oBACnF,IAAI,KAAK,YAAY,wBAAS;wBAAE,OAAO,IAAI,CAAC;oBAC5C,MAAM,KAAK,CAAC;gBACb,CAAC,CAAC,CAAC;gBACH,IAAI,CAAC,MAAM;oBAAE,OAAO,IAAI,CAAC;gBACzB,OAAO;oBACN,MAAM;oBACN,MAAM,EAAE,MAAM,IAAI,CAAC,WAAW,CAAC,UAAU,CAAC,OAAO,EAAE,SAAS,CAAC;oBAE7D,UAAU,EAAE,IAAA,sCAAkB,EAAC,MAAM,CAAC;iBACtC,CAAC;YACH,CAAC;SACD,CAAC;IACH,CAAC;IAUD,KAAK,CAAC,qBAAqB,CAAC,kBAA0B;QACrD,MAAM,MAAM,GAAG,IAAA,2CAA6B,EAAC,kBAAkB,CAAC,CAAC;QACjE,MAAM,OAAO,GAAG,MAAM,IAAI,CAAC,qBAAqB,CAAC,IAAI,CAAC;YACrD,MAAM,EAAE,EAAE,EAAE,EAAE,IAAI,EAAE;YACpB,KAAK,EAAE,EAAE,EAAE,EAAE,IAAA,cAAI,EAAC,GAAG,MAAM,GAAG,CAAC,EAAE;SACjC,CAAC,CAAC;QACH,KAAK,MAAM,EAAE,EAAE,EAAE,IAAI,OAAO,EAAE,CAAC;YAE9B,MAAM,MAAM,GAAG,IAAI,CAAC,SAAS,CAAC,iBAAiB,CAAC,EAAE,CAAC,KAAK,CAAC,MAAM,CAAC,MAAM,CAAC,CAAC,CAAC;YACzE,MAAM,MAAM,CAAC,sBAAsB,CAAC,EAAE,CAAC,CAAC;YACxC,MAAM,MAAM,CAAC,YAAY,CAAC,EAAE,CAAC,CAAC;QAC/B,CAAC;IACF,CAAC;CACD,CAAA;;kDApKY,uCAAuC;IADnD,IAAA,YAAO,GAAE;qCAGwB,8BAAa,EACN,6CAAoB,EAC/B,sBAAS,EACG,+CAAqB,EAC/B,yCAAkB,EAClB,yCAAkB;GAPrC,uCAAuC,CAoKnD"}