MCP Tools Reference

get_team_status

See who is online, what branch they're on, and what they're working on.

Parameters: None

Returns: List of team members with name, branch, working on, files today, last activity, online status

Ask your AI: "Who's online right now?" or "What branch is Sarah working on?"

get_recent_activity

Get recent coding activity from the team.

Parameters: author (optional), limit (default: 10), source (optional: git-hook, claude-code-cli), branch (optional)

Returns: List of turns with summaries, models, files touched, tools used, duration

Ask your AI: "What has the team been working on today?" or "Show me Marcus's recent activity"

get_file_history

See who last touched a specific file and what they did.

Parameters: file_path (required)

Returns: List of turns that touched that file

Ask your AI: "Who last modified auth.service.ts and why?"

search_turns

Full-text search across all captured coding sessions.

Parameters: query (required), branch (optional)

Returns: Matching turns with pagination

Ask your AI: "Search for discussions about rate limiting"

get_summary

Get the latest AI-generated daily or weekly summary of team activity.

Parameters: type (optional: daily, weekly; default: daily), branch (optional)

Returns: Summary with content, highlights, and time period

Ask your AI: "Give me this week's team summary"

get_knowledge_base

List AI-generated knowledge articles from your team's coding sessions.

Parameters: type (optional: daily, weekly, member), limit (default: 10, max: 100), branch (optional)

Returns: List of summaries with titles, content, highlights, and time periods

Ask your AI: "What does the knowledge base say about our API design?"

get_decisions

Get architectural and technical decisions made by the team.

Parameters: category (optional: architecture, technology, convention, security, infrastructure), limit (default: 10), branch (optional), projectId (optional)

Returns: Decisions with title, description, rationale, category, alternatives, affected files

Ask your AI: "What architecture decisions were made this week?" or "Why did we choose PostgreSQL?"

get_project_context

Get the full technical context document for a project — architecture, modules, stack, dependencies, patterns.

Parameters: project_name (optional; auto-detects from git remote if omitted)

Returns: Complete markdown document with technical overview

Ask your AI: "What's the architecture of this project?" or "Explain this project's tech stack"

detect_conflicts

Check if anyone else is working on overlapping files or systems.

Parameters: None

Returns: List of active conflicts with members and overlapping files

Ask your AI: "Are there any conflicts with what I'm working on?" or "Is anyone else editing the auth module?"

create_handoff

Create a handoff document for your current work so a teammate can pick up where you left off.

Parameters: title (required), whats_done (required), whats_left (required), gotchas (optional)

Ask your AI: "Create a handoff for my current work on the payment integration"

get_handoffs

List active handoffs for your team.

Parameters: status (optional: active, picked_up, completed)

Ask your AI: "Are there any handoffs waiting for me?"

pickup_handoff / complete_handoff

Pick up a handoff to start working on it, or mark one as completed.

Parameters: handoff_id (required)

Ask your AI: "Pick up the auth middleware handoff" or "Mark handoff xyz as complete"

generate_plan

Generate an AI-powered implementation plan with tasks, gap analysis, risks, and verification steps.

Parameters: title (required), description (required), project_id (optional), plan_id (optional — regenerate existing)

Ask your AI: "Create a plan for migrating the auth system to OAuth 2.0"

list_plans / get_plan / get_plan_markdown

List, view, or export plans.

Parameters: status (optional filter), limit (optional), plan_id (for detail/markdown)

Ask your AI: "Show me all in-progress plans" or "Get the full plan for the docs overhaul"

update_plan / update_plan_task / approve_plan

Update plan fields, change task statuses (auto-recalculates progress), or approve plans for execution.

Parameters: plan_id (required), task_id + status (for task updates), title, description, status (for plan updates)

Ask your AI: "Mark the auth task as completed" or "Approve the migration plan"

convert_plan_to_tasks

Convert a plan's tasks into individual sprint tasks, optionally assigning them to a sprint.

Parameters: plan_id (required), sprint_id (optional)

Ask your AI: "Convert the approved plan into sprint tasks"

create_task / update_task / list_tasks / get_task

Full task management — create, update, list, and view tasks with priority, labels, assignees, and sprint association.

Parameters: title, description, assignee_id, priority (low/medium/high/urgent), labels, project_id, sprint_id, status

Ask your AI: "Create a high-priority task for fixing the login bug" or "List all tasks assigned to me"

add_task_comment / link_session_to_task

Add comments to tasks or link coding sessions for full context tracking.

Parameters: task_id (required), content (for comments), session_id (for linking)

Ask your AI: "Add a note to the auth task: found the root cause in middleware"

create_sprint / list_sprints

Create and list sprints with goals and date ranges.

Parameters: title (required), goal (optional), start_date / end_date (ISO 8601), project_id

Ask your AI: "Create a 2-week sprint for the API migration"

list_agent_findings

List findings with filters by agent, project, severity, and status.

Parameters: agent_id (optional), project_id (optional), severity (optional: critical, high, medium, low, info), status (optional: open, dismissed, resolved), limit (default: 20)

Ask your AI: "Show me all critical security findings" or "List open findings for this project"

get_finding_detail / get_findings_stats

View full details of a finding or get aggregate stats by severity, status, agent, and project.

Parameters: finding_id (for detail)

Ask your AI: "How many critical findings do we have?" or "Show details of finding xyz"

dismiss_finding / resolve_finding

Dismiss a finding as not relevant, or mark it as resolved.

Parameters: finding_id (required)

Ask your AI: "Dismiss the N+1 query finding — it's already fixed"