Agent Claude Code
Explains code as a well-written narrative — from entry point to execution end, covering intent and design rather than line-by-line translation. Use when you need to understand how an unfamiliar system works, onboard someone to a codebase, or document complex execution flows.
Agent Claude Code
Critical thinking mode that challenges assumptions and probes reasoning. Use when you want a second opinion on an approach, need to stress-test a design decision, or want to be pushed to think deeper before committing to a solution.
Agent Claude Code
Systematic debugging agent. Use when tracking down a bug, diagnosing unexpected behavior, analyzing stack traces, or investigating a failing test. Follows a structured process from reproduction to root cause to verified fix.
Agent Claude Code
Planning agent that researches the codebase and produces a detailed, actionable implementation plan. Use when you need to think through an implementation before starting, outline multi-step tasks, or explore feasibility before writing code.
Agent Claude Code
Validates PR review comments for correctness, severity, and recommendation quality using full codebase context. Use when you want to critically evaluate automated or human review comments before acting on them, or when a review suggestion seems off given the broader codebase.
Agent Claude Code
Product management advisor. Use when creating GitHub issues, defining feature requirements, writing user stories, prioritizing work, or ensuring features align with user needs and business value.
Agent Claude Code
QA subagent for test planning, bug hunting, edge-case analysis, and implementation verification. Use when you need thorough test coverage, want to find bugs before shipping, or need to verify that a feature works correctly across edge cases.
Agent Claude Code
Security-focused code review specialist. Use when reviewing code for vulnerabilities, checking OWASP Top 10 compliance, auditing authentication/authorization, or reviewing AI/LLM integrations for security risks.
Agent Claude Code
Generates technical debt remediation plans for code, tests, and documentation. Use when you need to assess and prioritize technical debt, plan refactoring work, or create actionable remediation steps for code quality issues.
Command Claude Code
Release Spica as version $ARGUMENTS. Follow every step in order. Do not skip steps or proceed past a user-approval gate without confirmation.
Agent
Researches and outlines multi-step plans.
Agent
Meticulous QA subagent for test planning, bug hunting, edge-case analysis, and implementation verification.
Agent
Product management guidance for creating GitHub issues, aligning business value with user needs, and making data-driven product decisions.
Agent
Security-focused code review specialist with OWASP Top 10, Zero Trust, LLM security, and enterprise security standards.
Agent
Generate technical debt remediation plans for code, tests, and documentation.
Instructions file GitHub Copilot
Copilot instructions for spica-engine/spica, covering spica backend engine — copilot instructions, architecture overview, project layout, key path aliases (tsconfig.json) and module system & patterns.
Instructions file GitHub Copilot
Atomic Design methodology for organizing React components. Enforces strict rules for Atoms, Molecules, Organisms, Templates, and Pages — including what state, styling, and data-fetching each level is allowed.
Instructions file GitHub Copilot
The most comprehensive, practical, and engineer-authored performance optimization instructions for all languages, frameworks, and stacks. Covers frontend, backend, and database best practices with actionable guidance, scenario-based checklists, troubleshooting, and pro tips.
Instructions file GitHub Copilot
React + TypeScript + tRPC development standards for web app.
Instructions file
Claude Code instructions for spica-engine/spica, covering spica backend engine, architecture overview, project layout, key path aliases (tsconfig.json) and module system & patterns.
MCP server Claude CodeCodexCursor
MCP server for AI agents to interact with Spica servers — manage resources, databases, functions, storage, auditing, debugging and more. Runs locally from the @spica/mcp npm package.