Agent Claude Code
Enterprise audit logging for compliance (SOC 2, ISO 27001, HIPAA). Logs all AI operations, prompts, outputs, and file changes. Generates audit trails and compliance reports.
Agent Claude Code
Master orchestrator agent that coordinates the entire Claude Code workflow. Invoked via /boris command. Handles planning, delegation to specialist agents, verification loops, and learning from mistakes. Mirrors Boris Cherny's actual workflow.
Agent Claude Code
CI/CD feedback loop automation. Pushes code, monitors CI pipelines, parses failures, auto-fixes issues, and iterates until green. Integrates with GitHub Actions, GitLab CI, and other CI systems.
Agent Claude Code
Senior architect for design reviews, system design, architectural decisions, and technical planning. Use when planning significant changes, evaluating approaches, or reviewing design patterns.
Agent Claude Code
Simplify and improve code readability after changes WITHOUT changing functionality. Use after implementing features to clean up, reduce complexity, and improve maintainability.
Agent Claude Code
Generate and update documentation from code. Creates README sections, API docs, component docs, and keeps CLAUDE.md current.
Agent Claude Code
Intelligent git operations with safety features. Provides auto-commit with AI messages, undo capability, dirty file protection, checkpoint management, and attribution. Prevents accidental data loss.
Agent Claude Code
Native integration with GitHub Issues, Linear, and Jira. Fetches issue details, understands requirements, links commits to issues, and manages issue lifecycle. Use when working on tracked issues.
Agent Claude Code
Persistent memory system for cross-session context. Maintains project understanding, architectural decisions, progress state, and learned patterns. Invoke at session start/end or when context needs preservation.
Agent Claude Code
Mode system for behavioral segmentation. Switches between Architect, Code, Debug, and Review modes with different tool access and restrictions. Prevents accidents by limiting capabilities per mode.
Agent Claude Code
Debug production issues and incidents. Use when something is broken in production and needs rapid diagnosis and resolution.
Agent Claude Code
Automated code review for pull requests. Catches bugs, security issues, style problems, and suggests improvements. Use before merging any PR.
Agent Claude Code
Security scanning and vulnerability assessment. Runs SAST analysis, checks dependencies for CVEs, detects OWASP patterns, and ensures code meets security standards. Use before shipping or during code review.
Agent Claude Code
Generate comprehensive tests for code. Use when new features need tests, coverage is low, or critical paths need better testing.
Agent Claude Code
Comprehensive end-to-end application verification. Use before shipping features or merging PRs to ensure everything works correctly.
Command Claude Code
Invoke Boris, the master orchestrator, to handle any development task end-to-end. Boris will plan, coordinate specialists, verify, and ship.
Command Claude Code
Create a named checkpoint for easy rollback. Saves current state with a memorable name.
Command Claude Code
Push code, wait for CI, parse failures, auto-fix, and iterate until green. Automated CI feedback loop.
Command Claude Code
Complete git workflow - stage changes, commit with conventional message, push to remote, and create a pull request.
Command Claude Code
Show context window usage, Memory Bank status, and recommendations for context management.
Command Claude Code
Apply first principles thinking to break down complex problems and find optimal solutions.
Command Claude Code
Fetch a GitHub/Linear issue, understand requirements, implement the fix, and create a PR. End-to-end issue resolution.
Command Claude Code
Initialize Memory Bank for a new project. Creates memory structure and populates with project context.
Command Claude Code
Switch between operational modes (architect, code, debug, review, audit). Each mode has different tool access and behavioral constraints.