Package com.xebialabs.xlrelease.agentrules


package com.xebialabs.xlrelease.agentrules
  • Classes
    Class
    Description
    The agentic-triggers engine: matches every domain event against the user-authored agentic.Rule configurations and dispatches the configured AI agent for each match, posting the agent's analysis back where the event happened (a task comment, when the event carries a task) and recording every firing in the run history.
    The parsed "Then" half of an agent rule: author-defined outcome branches, each carrying an ordered action list.
    One action row: type, autonomy (auto vs propose), and templated parameters.
    One outcome branch: the author-chosen name and its ordered actions.
    One propose-mode action awaiting a human decision: the agent classified the event and the matched branch carries an action whose autonomy mode is propose, so instead of executing, the dispatcher parks it here for one-click approval on the Agent Rules page.
    Pending and recently decided propose-mode actions, newest first.
    One firing of an agent rule, as shown in the run-history view.
    Recent agent-rule runs, newest first, for the run-history view.
    The rules engine's OAuth2 identities, configured under xl.features.ai.agent-rules.oauth: Service identity (serviceToken()): short-lived access tokens via the client-credentials grant (RFC 6749 section 4.4) against the platform IdP.