td-review-session

Coordinate td review workflows to triage pending reviews and assign subagents.

3|1|Updated Dec 23, 2025
One-click install
npx skills add https://github.com/marcus/marcus-skills --skill td-review-session
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: td-review-session
Source: https://github.com/marcus/marcus-skills/tree/main/skills/td-review-session
Command: npx skills add https://github.com/marcus/marcus-skills --skill td-review-session

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Orchestrates a complete td review workflow: initializes review sessions, aggregates all pending items, and coordinates parallel subagents to perform detailed reviews, enabling faster triage and consistent follow-up actions.

Core Features & Use Cases

  • Session orchestration: Start and manage td review sessions, gather pending reviews, and maintain an auditable workflow.
  • Parallel triage: Spawn subagents to review items concurrently when there are independent reviews.
  • Automated classification: Classify reviews into obvious-fix, needs-tests, bigger-issue, clean, or unclear to drive appropriate actions.
  • Task creation and tracking: Create precise td tasks or bugs with contextual descriptions, attach to original reviews, and monitor progress.
  • State management and logging: Centralized state in td with structured logging to support traceability and recovery.

Quick Start

Start a new td review session and begin triaging all pending reviews.

Frequently Asked Questions about td-review-session

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I automate code review triage for a backlog with multiple independent items?

Automated code review triage orchestrates a review session to aggregate pending items and spawn parallel subagents for concurrent processing. This coordinates fast triage by classifying issues and assigning subagents to independent reviews simultaneously.

What's the best way to manage state and track tasks during a parallel code review?

Managing state during a parallel code review requires centralized tracking to ensure auditability and support recovery. The orchestrator maintains a deterministic workflow state, creating precise tasks with contextual descriptions and attaching them to original reviews.

How does automated review classification handle different types of code issues?

Automated review classification categorizes pending items into obvious-fix, needs-tests, bigger-issue, clean, or unclear. This classification drives appropriate follow-up actions by routing each category to specific task creation or subagent assignment workflows.

Can I process dependent code reviews in parallel without losing traceability?

Processing dependent code reviews in parallel is supported when reviews are independent. The orchestrator enforces deterministic behavior and parallelization where possible, maintaining structured logging and centralized state to ensure full traceability across the workflow.

When should I use a code review orchestrator instead of manual triage?

Use a code review orchestrator when managing a backlog with multiple independent items and dependencies between reviews. It provides automated classification, parallel processing, and structured state management that manual triage cannot efficiently handle at scale.

Does parallel code review work with existing task management and bug tracking workflows?

Parallel code review integrates with task management by creating precise tasks or bugs with contextual descriptions from classified review items. It attaches these tasks to original reviews and monitors progress within a centralized, auditable state management system.