Comments

Task comment management through MCP. Add, list, and update comments on tasks to facilitate collaboration and track decisions.

Open Source — GitScrum MCP Server is open source under the MIT license. Available on npm and GitHub. Model Context Protocol server for GitScrum — Claude, GitHub Copilot, Cursor, and any MCP-compatible client full operational access to your project management stack.

The comment tool provides 3 actions for managing task-level comments — the primary communication channel for task-specific collaboration. Every code review note, status update, implementation question, and technical decision that lives on a task flows through this tool. Comments support full Markdown, enabling rich content with code blocks, inline formatting, lists, and links.

Task comments differ from Discussions in scope. Discussions are project-level threads for broad topics and decisions. Comments are task-level — tightly scoped to a specific work item, forming the implementation narrative that team members reference during code reviews, standups, and retrospectives.

Actions Overview

Action	Purpose	Required Parameters
`list`	List all comments on a task	`taskuuid`, `company``slug`, `project_slug`
`create`	Add a new comment to a task	`taskuuid`, `company``slug`, `project_slug`, `comment`
`update`	Edit an existing comment	`uuid`, `taskuuid`, `company``slug`, `project_slug`, `comment`

Listing Comments

The list action returns all comments on a task in chronological order. Each comment includes its uuid, author, creation timestamp, and Markdown content. This gives you the full conversation history for any task.

Required Parameters

Parameter	Type	Description
`task_uuid`	string	Task UUID (from any task listing, `get`, or `by_code` response)
`company_slug`	string	Workspace identifier (from the `workspace` tool)
`project_slug`	string	Project identifier (from the `project` tool)

Example Prompts

You:  "Show all comments on task PROJ-42"
AI:   Resolves task by code → calls comment action=list → returns comment thread

You:  "What has been discussed on the login fix task?"
AI:   Finds task → calls comment action=list → summarizes the conversation

You:  "List comments on the API migration task"
AI:   Calls comment action=list → returns chronological comment history

Creating Comments

The create action adds a new comment to a task. Comments support full Markdown formatting — use code blocks for code snippets, bullet lists for action items, and headers for structured updates.

Required Parameters

Parameter	Type	Description
`task_uuid`	string	Task UUID to comment on
`company_slug`	string	Workspace identifier
`project_slug`	string	Project identifier
`comment`	string	Comment body in Markdown format

Example Prompts

You:  "Add a comment to PROJ-42: 'Fixed the null pointer exception
       in the auth middleware. See commit abc123.'"
AI:   Resolves task → calls comment action=create with Markdown comment

You:  "Comment on the login task with a code review note about
       the password validation regex"
AI:   Finds task → calls comment action=create with technical detail

You:  "Post a status update on the deployment task: staging tests passed,
       ready for production"
AI:   Calls comment action=create with status update formatted in Markdown

You:  "Add a comment listing the remaining work:
       - Update API docs
       - Add integration tests
       - Update changelog"
AI:   Calls comment action=create with a Markdown checklist

Markdown Support

Comments accept full Markdown. Common patterns for task comments:

## Status Update
Staging tests passed. All 247 assertions green.

## Remaining Work
- Update API documentation
- Add integration tests for edge cases
- Update CHANGELOG.md

## Code Reference
The fix is in `auth/middleware.ts`:

if (!token) { throw new UnauthorizedError('Token required'); }

Updating Comments

The update action replaces the content of an existing comment. This requires the comment's uuid (from a list response) plus the task context. The entire comment body is replaced with the new content.

Required Parameters

Parameter	Type	Description
`uuid`	string	Comment UUID (from `list` response)
`task_uuid`	string	Task UUID the comment belongs to
`company_slug`	string	Workspace identifier
`project_slug`	string	Project identifier
`comment`	string	New comment body in Markdown format (replaces existing content)

Example Prompts

You:  "Update my last comment on PROJ-42 to include the benchmark results"
AI:   Lists comments → identifies your latest → calls comment action=update
      with the revised Markdown content

You:  "Fix the typo in my comment about the API endpoint"
AI:   Lists comments → finds the comment → calls comment action=update

You:  "Replace my status update with: 'All tests passing. Deployed to staging.
       Production deploy scheduled for Monday.'"
AI:   Finds comment → calls comment action=update with new content

Task Collaboration Workflow

Comments are the real-time narrative of task execution. Combined with other MCP tools, they form a complete collaboration workflow:

1. Pick up the task

Check the task details and existing comments before starting work.

You:  "Show me task PROJ-42 and its comments"
AI:   Calls task action=by_code → then comment action=list
      → returns task details with full comment history

2. Post implementation notes

Leave comments as you work through the task — decisions made, blockers found, approaches tried.

You:  "Comment on PROJ-42: 'Investigated the timeout issue.
       Root cause is the missing connection pool limit.
       Implementing a fix with max 20 connections.'"
AI:   Calls comment action=create with the implementation note

3. Request feedback

Post code snippets or technical questions for team review.

You:  "Add a comment asking the team to review the error handling approach:
       should we retry on 503 or fail fast?"
AI:   Creates a comment with the question for team input

4. Update with progress

Keep the comment thread current as the task progresses.

You:  "Update my last comment — the connection pool fix is deployed to staging"
AI:   Finds your last comment → calls comment action=update

5. Document the resolution

Leave a final comment summarizing what was done before closing the task.

You:  "Add a closing comment on PROJ-42: 'Fixed by adding connection pool
       limit of 20. Deployed to production. Monitoring shows 0 timeouts
       since deploy.' Then complete the task."
AI:   Creates the comment → calls task action=complete

Comments vs Discussions

Choose the right tool for the conversation:

Aspect	Comments (this tool)	Discussions
Scope	Attached to a specific task	Project-level thread
Purpose	Implementation notes, code review, status updates	Architecture decisions, proposals, planning
Lifecycle	Lives with the task — moves with it, archives with it	Independent of individual tasks
Best for	"How is this task progressing?"	"What approach should we use?"

When a comment thread on a task grows into a broader decision, consider creating a Discussion to give the topic its own space. Reference the discussion in a task comment so the connection is clear.

Context Auto-Resolution

The comment tool supports automatic context resolution. If the AI assistant already knows the workspace, project, and task from a previous call in the conversation, it carries that context forward. You can say "add a comment" without re-specifying the task, as long as the task context was established earlier.

When a comment is created or listed, the MCP server returns context metadata (companyslug, projectslug, task_uuid) that the AI assistant uses for follow-up operations.

Next Steps

Discussions: Create project-level discussion threads for broader topics.
Tasks: Manage the tasks that comments are attached to.
Wiki: Document decisions and processes in project wiki pages.
Quick Start: Set up the MCP server if you haven't already.