From 325fe70ea3b48f6f0841921af76dee44fae232fc Mon Sep 17 00:00:00 2001 From: "copilot-swe-agent[bot]" <198982749+Copilot@users.noreply.github.com> Date: Fri, 6 Feb 2026 13:42:48 +0000 Subject: [PATCH 1/3] Initial plan From 26c45c76dc84f3d0f95268b04c68e9cbcca09470 Mon Sep 17 00:00:00 2001 From: "copilot-swe-agent[bot]" <198982749+Copilot@users.noreply.github.com> Date: Fri, 6 Feb 2026 13:50:52 +0000 Subject: [PATCH 2/3] Replace "Interrupt tool calls" with "Cancel tool calls" in AI Agent event handling docs Co-authored-by: maff <27403+maff@users.noreply.github.com> --- .../agentic-ai/aiagent/configuration/_event-handling.md | 8 ++++---- .../agentic-ai/aiagent/configuration/_event-handling.md | 8 ++++---- 2 files changed, 8 insertions(+), 8 deletions(-) diff --git a/docs/components/connectors/out-of-the-box-connectors/agentic-ai/aiagent/configuration/_event-handling.md b/docs/components/connectors/out-of-the-box-connectors/agentic-ai/aiagent/configuration/_event-handling.md index 8a567cd4ab3..a0eb249bc48 100644 --- a/docs/components/connectors/out-of-the-box-connectors/agentic-ai/aiagent/configuration/_event-handling.md +++ b/docs/components/connectors/out-of-the-box-connectors/agentic-ai/aiagent/configuration/_event-handling.md @@ -5,7 +5,7 @@ Configures how the AI Agent sub-process should behave when handling results from :::important In AI Agent sub-processes (implemented as ad-hoc sub-processes), only non-interrupting event subprocesses are supported. -As a result, even when using Interrupt tool calls, the event subprocess itself remains non-interrupting; cancellation happens at the tool-call level, not by stopping the parent sub-process. +As a result, even when using Cancel tool calls, the event subprocess itself remains non-interrupting; cancellation happens at the tool-call level, not by stopping the parent sub-process. See [How event subprocesses work with the AI Agent Sub-process](#how-event-subprocesses-work-with-the-ai-agent-sub-process) for more details. ::: @@ -13,7 +13,7 @@ See [How event subprocesses work with the AI Agent Sub-process](#how-event-subpr The **Event handling behavior** field can be set to one of these two options: - [Wait for tool call results](#wait-for-tool-call-results) (default). -- [Interrupt tool calls](#interrupt-tool-calls). +- [Cancel tool calls](#cancel-tool-calls). Consider the example scenario where the agent requested the execution of tools `A` and `B`, and tool `B` has already been completed, the following describes how each option behaves when an event is received. @@ -29,7 +29,7 @@ For the example scenario, the following sequence of messages would be sent to th 1. Tool B: `Tool B result`. 1. Event message: `Content from event message`. -#### Interrupt tool calls +#### Cancel tool calls When the configured event fires while one or more tool calls are still running, the agent: @@ -65,7 +65,7 @@ For the AI Agent Sub-process (implemented as an ad-hoc sub-process): - **Only non-interrupting event subprocesses are allowed** on the agent scope, by design and enforced by Modeler. - The **Event handling behavior** field controls how the agent treats running tool calls when such a (non-interrupting) event subprocess is triggered: - [Wait for tool call results](#wait-for-tool-call-results): the event handler runs only after all tool calls complete. - - [Interrupt tool calls](#interrupt-tool-calls): running tools are canceled, and their cancellation is surfaced to the LLM, even though the event subprocess itself remains non-interrupting from a BPMN perspective. + - [Cancel tool calls](#cancel-tool-calls): running tools are canceled, and their cancellation is surfaced to the LLM, even though the event subprocess itself remains non-interrupting from a BPMN perspective. By modeling an ad-hoc sub-process with an AI agent, you delegate control of sub-process execution (which tasks/tools run and when it completes) to the agent. diff --git a/versioned_docs/version-8.8/components/connectors/out-of-the-box-connectors/agentic-ai/aiagent/configuration/_event-handling.md b/versioned_docs/version-8.8/components/connectors/out-of-the-box-connectors/agentic-ai/aiagent/configuration/_event-handling.md index 5da8ceb4b34..3b4107a1de2 100644 --- a/versioned_docs/version-8.8/components/connectors/out-of-the-box-connectors/agentic-ai/aiagent/configuration/_event-handling.md +++ b/versioned_docs/version-8.8/components/connectors/out-of-the-box-connectors/agentic-ai/aiagent/configuration/_event-handling.md @@ -5,7 +5,7 @@ Configures how the AI Agent sub-process should behave when handling results from :::important In AI Agent sub-processes (implemented as ad-hoc sub-processes), only non-interrupting event subprocesses are supported. -As a result, even when using Interrupt tool calls, the event subprocess itself remains non-interrupting; cancellation happens at the tool-call level, not by stopping the parent sub-process. +As a result, even when using Cancel tool calls, the event subprocess itself remains non-interrupting; cancellation happens at the tool-call level, not by stopping the parent sub-process. See [How event subprocesses work with the AI Agent Sub-process](#how-event-subprocesses-work-with-the-ai-agent-sub-process) for more details. ::: @@ -13,7 +13,7 @@ See [How event subprocesses work with the AI Agent Sub-process](#how-event-subpr The **Event handling behavior** field can be set to one of these two options: - [Wait for tool call results](#wait-for-tool-call-results) (default). -- [Interrupt tool calls](#interrupt-tool-calls). +- [Cancel tool calls](#cancel-tool-calls). Consider the example scenario where the agent requested the execution of tools `A` and `B`, and tool `B` has already been completed, the following describes how each option behaves when an event is received. @@ -29,7 +29,7 @@ For the example scenario, the following sequence of messages would be sent to th 1. Tool B: `Tool B result`. 1. Event message: `Content from event message`. -#### Interrupt tool calls +#### Cancel tool calls When the configured event fires while one or more tool calls are still running, the agent: @@ -65,7 +65,7 @@ For the AI Agent Sub-process (implemented as an ad-hoc sub-process): - **Only non-interrupting event subprocesses are allowed** on the agent scope, by design and enforced by Modeler. - The **Event handling behavior** field controls how the agent treats running tool calls when such a (non-interrupting) event subprocess is triggered: - [Wait for tool call results](#wait-for-tool-call-results): the event handler runs only after all tool calls complete. - - [Interrupt tool calls](#interrupt-tool-calls): running tools are canceled, and their cancellation is surfaced to the LLM, even though the event subprocess itself remains non-interrupting from a BPMN perspective. + - [Cancel tool calls](#cancel-tool-calls): running tools are canceled, and their cancellation is surfaced to the LLM, even though the event subprocess itself remains non-interrupting from a BPMN perspective. By modeling an ad-hoc sub-process with an AI agent, you delegate control of sub-process execution (which tasks/tools run and when it completes) to the agent. From d7e0910088bce90f84dafa5b9dc462f12222137a Mon Sep 17 00:00:00 2001 From: "copilot-swe-agent[bot]" <198982749+Copilot@users.noreply.github.com> Date: Fri, 6 Feb 2026 13:57:47 +0000 Subject: [PATCH 3/3] docs: format "Cancel tool calls" with bold and use "cancellation" for consistency Co-authored-by: maff <27403+maff@users.noreply.github.com> --- .../agentic-ai/aiagent/configuration/_event-handling.md | 4 ++-- .../agentic-ai/aiagent/configuration/_event-handling.md | 4 ++-- 2 files changed, 4 insertions(+), 4 deletions(-) diff --git a/docs/components/connectors/out-of-the-box-connectors/agentic-ai/aiagent/configuration/_event-handling.md b/docs/components/connectors/out-of-the-box-connectors/agentic-ai/aiagent/configuration/_event-handling.md index a0eb249bc48..3de6d93cf6c 100644 --- a/docs/components/connectors/out-of-the-box-connectors/agentic-ai/aiagent/configuration/_event-handling.md +++ b/docs/components/connectors/out-of-the-box-connectors/agentic-ai/aiagent/configuration/_event-handling.md @@ -5,7 +5,7 @@ Configures how the AI Agent sub-process should behave when handling results from :::important In AI Agent sub-processes (implemented as ad-hoc sub-processes), only non-interrupting event subprocesses are supported. -As a result, even when using Cancel tool calls, the event subprocess itself remains non-interrupting; cancellation happens at the tool-call level, not by stopping the parent sub-process. +As a result, even when using **Cancel tool calls**, the event subprocess itself remains non-interrupting; cancellation happens at the tool-call level, not by stopping the parent sub-process. See [How event subprocesses work with the AI Agent Sub-process](#how-event-subprocesses-work-with-the-ai-agent-sub-process) for more details. ::: @@ -38,7 +38,7 @@ When the configured event fires while one or more tool calls are still running, - Keeps the surrounding **ad-hoc sub-process scope active**: the agent decides, based on the new LLM response, which elements to activate next or whether to complete the sub-process. :::note -No BPMN interrupting event subprocess is involved; the interruption happens purely in the agent’s control flow (job worker), not by canceling the BPMN parent scope. +No BPMN interrupting event subprocess is involved; the cancellation happens purely in the agent’s control flow (job worker), not by canceling the BPMN parent scope. ::: For the example scenario, the following sequence of messages would be sent to the LLM: diff --git a/versioned_docs/version-8.8/components/connectors/out-of-the-box-connectors/agentic-ai/aiagent/configuration/_event-handling.md b/versioned_docs/version-8.8/components/connectors/out-of-the-box-connectors/agentic-ai/aiagent/configuration/_event-handling.md index 3b4107a1de2..4e06e80c1ee 100644 --- a/versioned_docs/version-8.8/components/connectors/out-of-the-box-connectors/agentic-ai/aiagent/configuration/_event-handling.md +++ b/versioned_docs/version-8.8/components/connectors/out-of-the-box-connectors/agentic-ai/aiagent/configuration/_event-handling.md @@ -5,7 +5,7 @@ Configures how the AI Agent sub-process should behave when handling results from :::important In AI Agent sub-processes (implemented as ad-hoc sub-processes), only non-interrupting event subprocesses are supported. -As a result, even when using Cancel tool calls, the event subprocess itself remains non-interrupting; cancellation happens at the tool-call level, not by stopping the parent sub-process. +As a result, even when using **Cancel tool calls**, the event subprocess itself remains non-interrupting; cancellation happens at the tool-call level, not by stopping the parent sub-process. See [How event subprocesses work with the AI Agent Sub-process](#how-event-subprocesses-work-with-the-ai-agent-sub-process) for more details. ::: @@ -38,7 +38,7 @@ When the configured event fires while one or more tool calls are still running, - Keeps the surrounding **ad-hoc sub-process scope active**: The agent decides, based on the new LLM response, which elements to activate next or whether to complete the sub-process. :::note -No BPMN interrupting event subprocess is involved; the interruption happens purely in the agent’s control flow (job worker), not by canceling the BPMN parent scope. +No BPMN interrupting event subprocess is involved; the cancellation happens purely in the agent’s control flow (job worker), not by canceling the BPMN parent scope. ::: For the example scenario, the following sequence of messages would be sent to the LLM: