Release documenter

Release documenter

Instructions

• Should always use Atlassian MCP server to perform actions on confluence
  • If not authenticated with MCP server ask user to Authenticate
• workspace should be determined or provided
  • It is the directory that holds all cloned repositories (system's) will most likely be a parent directory from where claude has been called from, can be verified by finding the .git folder and taking this at the root directory of the system, meaning the parent directory will be the parent of that folder
  • If cannot be determined, ask the user to provide the path
  • Should store this knowledge for future prompts if found to be correct, if ever becomes incorrect (folder not found) ask user again.
• If user has asked to make a code release, afterwards ask if they want to make a release document for the version that was just release
  • If version not provided, and the system they want to make a release version for is cloned (within the workspace directory), fetch latest master branch version and latest tag version from origin and use tag version as version
  • If they say no then exit
• If yes or was explicitly asked to make a release document and wasn't necessarily after a code release then proceed.
• version follows semantic versioning
• Ask the user if they want to use a specific page (Provided as a URL) as a template for the release document.
  • Have a default value of: https://siteminder.atlassian.net/wiki/spaces/AProg/pages/4311810049/v+version
  • Asks the user if they want to provide a template for the release document (can be remembered for later and mapped differently for each system, saved for future prompts)
  • The template provided may have variables defined in the format [[]]
  • When replacing values look for [[system]], [[version]], [[deploymentDate]], [[userMakingRelease]]
• For the system the release document is being made for, ask the user to provide the parent folder that holds the systems releases manually
  • If provided then continue, if asked to try find it then use the MCP server to find the location where release documentation is kept for that system
    • page is normally labeled as v<version>
• Use the template to create a new release document on the new version specified and update all relevant fields in the document to reflect the new version number
  • Be sure to not alter any formatting / layout of the original template being used
• If there any any jql queries (Used by Jira Legacy / Insert Jira Issue modal) it should be updated also with the relevant fixVersion (version) and project (system)
• If PVT tests are provided in a table, check diff between tag of version provided and the tag before and populate the PVT table appropriate, if not found then keep empty PVT table.
• If there is a checklist on the release page, tick what's appropriate; noting that if any relate to a user validating or verifying services are up and running keep unchecked.
• Output the link to the new page along with all relevant information

Examples

User provides, "Make a release document for demand-manager"

• system would be demand-manager
• using workspace location, fetch latest git changes for the system on master branch and determine the latest tag
• Ask the user if tag version found is the version that they want to make the release document for
• If not then let user provide version
• Asks the user if they want to provide a template for the release document
• Use Atlassian MCP server to determine where the release documents for the system are kept
  • if not found ask user to provide URL to parent directory that holds all release document pages
• Uses Atlassian MCP server to create the release document by duplicating the release document template in the same directory but renaming the page name to be the version we're making the release document for.
• Updates relevant details of new release document correctly

User provides, "Make a v1.0.0 release document for demand-manager"

• system would be demand-manager
• version would be v1.0.0
• As above