~50 tools return hardcoded mock data

[davidefiocco]

Hey @RyanCardin15 thanks again for this useful tool, I'm spreading the word about it!

One issue that may discourage adoption is that the MCP server currently features very many tools that are not implemented yet, and this can clutter context, increase token usage, etc.

I appreciate that the newly introduced ALLOWED_TOOLS env var can work around this, if set e.g. to

ALLOWED_TOOLS=listWorkItems,getWorkItemById,searchWorkItems,getRecentlyUpdatedWorkItems,getMyWorkItems,createWorkItem,updateWorkItem,addWorkItemComment,updateWorkItemState,assignWorkItem,createLink,bulkCreateWorkItems,getBoards,getBoardColumns,getBoardItems,moveCardOnBoard,getSprints,getCurrentSprint,getSprintWorkItems,getSprintCapacity,getTeamMembers,listProjects,getProjectDetails,createProject,getAreas,getIterations,createArea,createIteration,getProcesses,getWorkItemTypes,getWorkItemTypeFields,listRepositories,getRepository,createRepository,listBranches,searchCode,browseRepository,getFileContent,getCommitHistory,listPullRequests,createPullRequest,getPullRequest,getPullRequestChangedFiles,getPullRequestComments,approvePullRequest,mergePullRequest

Still, it would be helpful to either document this list in the README explicitly or remove them and reintroduce them properly in the future. Not activating the mock tools looks to me as a more sensible default, because the above var clutters mcp.json considerably. My 2 cents. Also glad to file a PR if a choice is made in this direction.

I think it would be worth cosidering this also given #22 (comment) and #10 (comment) which both mention this problem in some form.

Thanks for your great work on this!