Install in seconds

Install this skill

Copy the command and run it in your terminal. You can review the source before installing.

terminal

git clone https://github.com/a5c-ai/babysitter

Works with Git. The repository opens in your current directory.

🔍

QualityJavaScript

Code Review

Performs automated code review across correctness, security, performance, maintainability, and test coverage. Generates scores, findings, and actionable improvement recommendations.

1.6K stars94 forksAdded 2026/07/20

View source on GitHub Download ZIP

agent-orchestrationagent-skillsagentic-aiagentic-workflowai-agentsai-automationbabysitterclaude-codeclaude-code-skillsclaude-code-workflowsclaude-skillsclaude-workflowscodex-plugincodex-skillscodex-workflowhermes-pluginpi-extensiontrustworthy-aivibe-coding

Documentation

README

From the original repository

Code Review

Multi-dimensional code review across correctness, security, performance, maintainability, and test coverage.

Agents

Amelia (Developer) - bmad-dev-amelia (full team mode)
Barry (Solo Dev) - bmad-solodev-barry (quick flow mode)

Workflow

Review correctness: logic, edge cases, error handling
Review security: validation, auth, data protection
Review performance: efficiency, resources, caching
Review maintainability: clarity, naming, SOLID
Review test coverage: missing tests, quality
Rate overall quality and identify tech debt
Generate actionable recommendations

Inputs

projectName - Project name
storyResults - Implementation results to review

Outputs

Overall rating and score (0-100)
Category scores (5 dimensions)
Finding list with severity and suggestions
Technical debt identification
Action items

Process Files

bmad-orchestrator.js - Phase 4 code review
bmad-implementation.js - Sprint code review
bmad-quick-flow.js - Quick flow review

More from a5c-ai

Other Claude Code skills by this author in the directory.

Assimilate Popular Workflows

Searches public GitHub repositories for SKILL.md files, classifies repos by archetype, and extracts transferable processes and plugin ideas. Use when researching external AI agent workflows or finding extractable procedural patterns.

Babysit Babysitter Issues

Fetches open issues assigned to a specific agent and orchestrates their resolution sequentially using predefined skills. Useful for automated issue processing in agentic workflows.

Catalog Babysitter Users

Discover public codebases that import defineTask from @a5c-ai/babysitter-sdk and maintain a deduplicated catalog of those projects. Excludes any project named "babysitter". Run this to find, catalog, or refresh lists of projects using babysitter.

Fix Failing Pipelines

Automatically identify and fix failing CI workflows on the staging branch using the babysitter agent. Diagnoses the most recent pipeline failure, invokes an AI agent to apply repairs directly to the staging branch, and verifies the fix.

Process Builder

Scaffold new process definitions for the babysitter event-sourced orchestration framework. Guides a 3-phase workflow from research to implementation, following SDK patterns and best practices.

Retrospect External Babysitter Run

Analyzes a babysitter run in an external public repository, producing a retrospective with concrete suggestions for process improvement. Use when reviewing how someone else used a babysitter process or when examining committed run logs.

More Quality skills

Code Review is one of 164 Quality skills cataloged on DirSkills.

Browse all Quality skills →

007 — License to Audit

Performs comprehensive security audits, threat modeling (STRIDE/PASTA), code reviews, and incident response across code, infrastructure, APIs, and AI agents. Use for red/blue team exercises, OWASP compliance, and hardening any project.

Accessibility Audit

Conduct accessibility audits for WCAG compliance, identify barriers, and get remediation guidance to ensure inclusive digital experiences.

AccessLint Audit

Find and fix WCAG 2.2 accessibility issues with two modes: report (sweep a codebase or page for a prioritized report) and fix (audit, edit, verify loop). Prefers live-DOM auditing, falls back to browser-MCP or HTML strings.