Story Solution

Designs the technical solution for a story: reads the story context, the answered questions, and the BA/SA/VD parent context, then produces the solution write-up and architecture diagrams.

By default the result is written to the ticket’s Solution and Diagrams fields and the story moves to Ready For Development. When contentOutput targets Confluence, the solution is instead published as a Confluence page (with the diagram as a Mermaid section), reviewer inline comments on that page are read before writing, and answers are posted back as comment replies.

Parameters

Configured via customParams in the agent JSON or the project .dmtools/config.js.

• solutionField — tracker field that receives the solution text. Default: Solution.
• diagramField — tracker field that receives the diagram. Set to an empty string to prepend the diagram to the solution as a code block instead. Default: Diagrams.
• outputType — replace (default) overwrites the solution field; append keeps the existing field content and appends the new solution after a separator.
• requireDiagram — when true, the run fails if the agent did not produce outputs/diagram.md. Default: false.
• checkOpenPR — when true, skip tickets that already have an open solution PR. Default: false.
• autoStartDevelopment / autoStartDevelopmentConfigFile — after a successful solution, automatically trigger the development workflow for the ticket using the given agent config file.
• contentOutput — output routing block; see below.

contentOutput

• contentOutput.target — jira_field (default, current behavior), confluence, or both.
• contentOutput.space / contentOutput.parentPageId — where the ticket’s solution page lives (required for Confluence targets).
• contentOutput.pageTitleSuffix — appended to the page title, e.g. Solution Design.
• contentOutput.includeInlineComments — fetch the page’s inline comments into the input folder before the run and publish outputs/confluence_replies.json after it. Default: true.
• contentOutput.updateTrackerField — for target: confluence, also write the page link into the solution field. Default: true.

Downstream consumers: other agents in the pipeline (for example the affected-repos / task-splitting flows that parse the Solution field) read the tracker field, not Confluence. When such a consumer exists, use target: 'both' — the field keeps the full machine-readable content and the Confluence page becomes the human-facing mirror. target: 'confluence' alone would leave only a link in the field and break those agents.

Human doc: agents/docs/agents/story_solution.md

Attributes

Prompt snapshot

Full prompt / instruction set: agents/snapshots/story_solution.md

Actions

preJSAction: checkWipLabel.js

Check WIP Label Pre-Action

• Source: agents/js/checkWipLabel.js
• Side effects:
  • post Jira comment
• Parameters (customParams):
  • checkOpenPR

preCliJSAction: preCliSolutionSetup.js

Pre-CLI Solution Setup

• Source: agents/js/preCliSolutionSetup.js
• No detected file I/O or side effects.

postJSAction: writeSolutionAndDiagrams.js

Write Solution and Diagrams Post-Action

• Source: agents/js/writeSolutionAndDiagrams.js
• Side effects:
  • post Jira comment
  • assign ticket
• Parameters (customParams):
  • autoStartDevelopment
  • autoStartDevelopmentConfigFile
  • contentOutput
  • diagramField
  • outputType
  • requireDiagram
  • solutionField

LLM step

• outputType: none
• Prompt snapshot: agents/snapshots/story_solution.md

Custom params

• checkOpenPR (used by JS action)
• autoStartDevelopment (used by JS action)
• autoStartDevelopmentConfigFile (used by JS action)
• contentOutput (used by JS action)
• diagramField (used by JS action)
• outputType (used by JS action)
• requireDiagram (used by JS action)
• solutionField (used by JS action)

Generated by js/agentDocGenerator.js