Agile plugin and AI Agent integration

• What the AI gets
• MCP tools and AI assistant tools (PRO)
• Tool arguments

This article describes how the Agile plugin works with the Redmine Agent plugin. Everything below needs the Agent plugin installed; without it Agile works as usual. All data and actions follow the permissions of the user and, for MCP, the scopes of the OAuth token. See MCP server and AI assistant tools for general setup.

The Agile module must be enabled in the project for sprints and issue fields.

What the AI gets

• Issue answers: get_issue, create_issue and update_issue return an agile object: the sprint (id, name), story_points and the card color (when colors are enabled).
• New arguments of create_issue and update_issue (when the user may add or edit issues in a project with the Agile module): story_points (integer), agile_sprint_id (0 removes the issue from its sprint), agile_color (a card color, empty string clears).

MCP tools and AI assistant tools (PRO)

Tool	Needs	Description
list_agile_sprints	View issues	sprints of a project: id, name, status (open, active, closed), dates and sharing; sprints overlapping from to to. Also available to the built-in assistant
get_issues_timeline	View issues	what happened to up to 100 issues in a window: status changes, assignees, sprint entries, durations in statuses; modes events and durations, sections chosen with include, compact for a dense encoding. Also available to the built-in assistant
manage_agile_sprint	Manage sprints	action: create, update, delete a sprint (deleting unlinks it from its issues)

Tool arguments

проверка

list_agile_sprints: sprints of a project. Same rules as the sprints REST API.

Argument	Description
project	required; project id or identifier
name	part of the sprint name
status	open, active or closed
from, to	dates; sprints overlapping the interval
include_shared	add sprints shared into the project
limit, offset	paging; limit 0-100, default 25, 0 returns only total_count
fields	columns to return; -name excludes one
compact	columns once and rows as arrays

get_issues_timeline: history of issues. The answer is the same as in REST API issues timeline, which also shows how to calculate cycle time, lead time and velocity from it.

Argument	Description
issue_ids	up to 100 issue ids. Instead, pass fields with agile_sprint_id and values to take every issue that was in a sprint
fields	status, assignee, sprint, story points, estimates, done ratio, version, tracker, priority, category, parent, dates; "all" or a list of values
from, to	window, UTC dates
mode	events or durations
count_only	only the total and the number of events per issue
include	sections of the answer
limit, offset	paging over issues; default 20, maximum 100
compact	dense encoding: schema once, records as arrays

manage_agile_sprint: create, update or delete a sprint. Validation errors (overlapping dates, a sprint with open issues cannot be closed) come back as the tool error.

Argument	Description
action	required; create, update or delete
project	project id or identifier, for create
sprint_id	for update and delete
name, description	sprint name and description
status	open, active or closed
sharing	none, descendants, hierarchy, tree or system
start_date, end_date	sprint dates

Example prompts for a client: "move all open issues assigned to me in project Web to the active sprint", "how long did the issues of the last sprint stay in In progress?".

The same data is available without AI: REST API Agile Sprint and REST API issues timeline.