← Documentation

MCP Tools Reference

17 consolidated tools covering 70+ operations for managing your entire project lifecycle. Works with Claude Code, Claude Desktop, Cursor, or any MCP-compatible agent.

Consolidated Tool Architecture

Instead of exposing 60+ individual tools, Sprintra uses a method dispatch pattern. Each tool groups related operations under a single name with a method parameter that selects the specific action. This means:

  • Fewer tools for the AI to discover and understand
  • Lower token cost per interaction (smaller tool schemas)
  • Compatible with all MCP clients including Claude Desktop's tool limits
  • Same full functionality — nothing was removed, just organized better

manage_projects

4 methods

Create, update, find, and configure projects. Each project maps to a code repository.

MethodDescriptionParameters
create

Create a new project linked to a repository

name, description, tech_stack, repo_path

update

Update project details, status, or description

project_id, name?, description?, status?, tech_stack?

resolve

Find a project by name or repo path

name?, repo_path?

generate_claude_md

Generate a CLAUDE.md file for auto-routing

project_id

Example Prompt

"Create a Sprintra project for this repo with Next.js, TypeScript, and Prisma"

Creates a project linked to your current directory with tech stack metadata. Future MCP calls from this directory auto-target this project.

manage_features

3 methods

Create and update epic-level work items with acceptance criteria, priority, and phase tracking.

MethodDescriptionParameters
create

Create an epic/feature with acceptance criteria

project_id, title, description, priority, phase?, acceptance_criteria?

update

Update status, priority, or details

feature_id, title?, status?, priority?, description?

get_bundle

Get feature with all stories and dependencies

feature_id

Example Prompt

"Create a feature for user authentication with high priority and acceptance criteria"

Creates a feature with completeness scoring. The response includes a score showing how well-defined the feature is.

manage_stories

3 methods

Create, update, and batch-update tasks within features. Stories are the work items your team executes.

MethodDescriptionParameters
create

Create a task/story under a feature

feature_id, title, description, type (story|task|bug|chore), story_points?

update

Update story status or details

story_id, title?, status?, description?, story_points?

batch_update

Update multiple stories at once

updates[] (array of story updates)

Example Prompt

"Create stories for login, signup, and password reset under the auth feature"

Creates 3 stories underneath a feature. Each entity gets a completeness score showing how well-defined it is.

manage_sprints

3 methods

Create and manage time-boxed iterations. Assign stories to sprints and track progress.

MethodDescriptionParameters
create

Create a new sprint with goal and dates

project_id, name, goal, start_date, end_date

update

Update sprint status, goal, or dates

sprint_id, name?, goal?, status?, start_date?, end_date?

assign

Assign a feature or story to a sprint

sprint_id, feature_id?, story_id?

Example Prompt

"Create a 2-week sprint starting Monday focused on the authentication feature"

Creates a sprint, assigns relevant features/stories, and sets the goal. When the first story starts, the sprint auto-activates.

manage_decisions

4 methods

Architecture Decision Records (ADRs). Capture the why behind technical choices for long-term context.

MethodDescriptionParameters
add

Record a technical decision with context and consequences

project_id, title, context, decision, consequences, category

supersede

Replace a decision with a newer one

decision_id, new_decision_id, rationale

get_conflicts

Find potentially conflicting decisions

project_id

resolve_conflict

Resolve a flagged conflict between decisions

project_id, conflict_id, resolution

Example Prompt

"Record a decision: we're using PostgreSQL instead of MongoDB for strong consistency"

Creates an ADR with context, decision, and consequences. Decisions persist across AI sessions and are surfaced when relevant.

manage_documents

5 methods

Knowledge base documents with versioning — PRDs, specs, runbooks, and more.

MethodDescriptionParameters
create

Create a KB document with optional template

project_id, title, content, doc_type, template?, tags?

update

Update document content (versioned)

document_id, title?, content?, status?

get

Get full document with content

document_id

get_versions

List version history

document_id

restore_version

Restore a previous version

document_id, version

Example Prompt

"Create a technical spec for the payment integration covering Stripe webhooks and error handling"

Creates a versioned KB document. Updates are tracked with full version history you can browse and restore.

manage_notes

1 methods

Quick-capture notes for ideas, observations, and context that don't fit elsewhere.

MethodDescriptionParameters
add

Quick-capture a note

project_id, content, tags?, feature_id?, source?

Example Prompt

"Note: the Stripe test API has a 100 req/s rate limit we need to account for"

Captures a quick note tagged for future reference. Notes can be linked to features for additional context.

manage_git

5 methods

Sync git history, link commits to features, and track AI-authored vs human-authored code.

MethodDescriptionParameters
sync_commits

Import recent git commits into Sprintra

project_id, limit?

link_commits

Link commits to a feature by hash

project_id, feature_id, commit_hashes[]

unlink_commits

Remove commit-feature links

project_id, commit_hashes[]

get_status

Get current repo git status

project_id

get_feature_stats

Git stats per feature (commits, files, AI %)

project_id

Example Prompt

"Sync the last 20 git commits and link them to the authentication feature"

Imports commit metadata and associates them with features. The dashboard shows AI authorship percentage and commit timeline per feature.

manage_sessions

3 methods

Brainstorm sessions with AI personas — technical architect, product manager, skeptical reviewer, and more.

MethodDescriptionParameters
start

Start a brainstorm session with a persona

project_id, persona

log_message

Log a message in the active session

session_id, role, content

complete

End session with summary and outcomes

session_id, summary, decisions_made?, features_created?

Example Prompt

"Start a brainstorm session as a technical architect to design the API gateway"

Creates a session with the architect persona. Key points are logged, and the session summary captures decisions and features created.

manage_work_sessions

3 methods

Track coding sessions for audit trail and session continuity across conversations.

MethodDescriptionParameters
start

Start tracking a coding session

project_id, goal?

end

End work session with summary

session_id, summary, files_changed?, stories_completed?

get_active

Check for an active work session

project_id

Example Prompt

"Start a work session — I'm implementing the payment webhook handler"

Tracks what was accomplished, files changed, and stories completed. Provides continuity when resuming work in a new conversation.

manage_dependencies

3 methods

Track relationships between features — which features block others.

MethodDescriptionParameters
add

Mark one feature as depending on another

project_id, feature_id, depends_on_id, dep_type?

remove

Remove a dependency link

project_id, feature_id, depends_on_id

get_graph

Get full dependency graph

project_id

Example Prompt

"Mark the payment feature as depending on user authentication"

Creates a dependency link. The dashboard visualizes the graph and flags blocked features in sprint planning.

manage_criteria

4 methods

Define, track, and verify acceptance criteria for features.

MethodDescriptionParameters
add

Add an acceptance criterion to a feature

feature_id, description, category?

remove

Remove a criterion

feature_id, criterion_id

update_status

Mark criteria as met or not met

feature_id, criterion_id, status

verify

AI-verify all criteria for a feature

feature_id

Example Prompt

"Add acceptance criteria to the auth feature: users can login with email, OAuth works with Google"

Adds structured criteria that can be individually tracked. The completeness indicator shows progress across all criteria.

manage_releases

3 methods

Track release milestones and auto-generate release notes from completed work.

MethodDescriptionParameters
create

Create a release milestone

project_id, name, version, target_date?, description?

update

Update release details or status

release_id, name?, version?, status?, target_date?

generate_notes

Auto-generate release notes from completed features

release_id

Example Prompt

"Create a v1.0 release targeting May 1st with the auth and payment features"

Creates a release milestone. When features are done, auto-generate release notes from the completed work.

manage_comments

4 methods

Add threaded comments to any entity — stories, features, decisions, and more.

MethodDescriptionParameters
add

Add a comment to an entity

entity_type, entity_id, content, author?

list

List comments on an entity

entity_type, entity_id

update

Edit a comment

comment_id, content

delete

Remove a comment

comment_id

Example Prompt

"Add a comment to story s-abc123: Started implementation, touching auth.ts and middleware.ts"

Adds audit trail to any work item. Comments capture approach, decisions, and context for future reference.

manage_custom_fields

3 methods

Define and set custom fields on any entity for project-specific metadata.

MethodDescriptionParameters
define

Define a custom field schema

project_id, entity_type, field_name, field_type, options?

set

Set custom field values on an entity

entity_type, entity_id, fields

get

Get custom field values for an entity

entity_type, entity_id

Example Prompt

"Add a custom field 'complexity' with options low/medium/high to all features"

Extends the data model without code changes. Custom fields appear on entities and can be used for filtering and reporting.

read_data

17 methods

Read-only queries for listing and viewing project data — projects, features, stories, sprints, decisions, documents, activity, and more.

MethodDescriptionParameters
list_projects

List all projects in your organization

get_project_context

Get project summary with feature/sprint/story counts

project_id

list_features

List all features with optional status filter

project_id, status?

list_stories

List stories, optionally filtered by feature

project_id, feature_id?, status?

get_current_sprint

Get the currently active sprint

project_id

list_sprints

List all sprints for a project

project_id

list_decisions

List all decisions, optionally by category

project_id, category?, status?

list_documents

List all documents for a project

project_id, doc_type?

list_notes

List all notes for a project

project_id

list_releases

List all releases

project_id

list_sessions

List brainstorm sessions

project_id

list_work_sessions

List work sessions

project_id

list_activity

Recent activity feed

project_id, limit?

list_traces

Execution traces

project_id

get_ai_guidance

Get AI-generated guidance for current project state

project_id, topic?

get_next_work

Recommended next story to work on

project_id

get_context_snapshot

Restore context from a previous snapshot

project_id

Example Prompt

"What should I work on next?" or "Show me all in-progress features"

Read-only queries that surface project state. The agent uses these to understand priorities, check progress, and recommend next actions.

Agent API Endpoints

In addition to MCP tools, Sprintra provides two optimized REST endpoints that replace 5-7 MCP tool calls with a single HTTP request. These are designed for AI agents that prefer direct API access.

GET /api/agent/context?project_id=...

Full project snapshot in one call — current sprint with progress, in-progress stories, recent decisions, active blockers, last work session, activity feed, and stats. Response time ~20ms.

GET /api/agent/next-work?project_id=...

Recommended next story with full context — feature details, related decisions, dependencies, and a readiness score. Replaces manual prioritization across multiple queries.

Common Workflows

Starting a new coding session

  1. Agent calls manage_work_sessionsget_active to check for an existing session
  2. If none, calls manage_work_sessionsstart
  3. Calls read_dataget_next_work for the recommended next task
  4. Updates story status to in_progress via manage_storiesupdate

Ending a coding session

  1. Agent calls manage_work_sessionsend with summary
  2. Calls manage_gitsync_commits to import new commits
  3. Calls manage_gitlink_commits to associate commits with features

Planning a sprint

  1. Review backlog with read_datalist_features / list_stories
  2. Create sprint with manage_sprintscreate
  3. Assign work with manage_sprintsassign
  4. Check blockers with manage_dependenciesget_graph