skills

Multi-axis code review across correctness, readability, architecture, security, and performance, applying irega project patterns. Use when reviewing code before merge, evaluating agent-generated output, or doing a quality pass on any code change. Use when user says "code review", "review this", "review my changes", or invokes /irega-code-review.

Specification-driven feature planning with OpenSpec. Use when starting a non-trivial feature, designing a system change, or when "what to build" isn't settled yet. Triggers on phrases like "plan this", "let's spec this out", "before we build", or when the scope is unclear.

Security-focused review from an OWASP/threat-modeling angle, covering input validation, auth, data protection, infrastructure, and third-party integrations. Works on any artifact (file, component, PR diff, or described architecture), not just open PRs. Use when user asks for a security review, security audit, or wants to check vulnerabilities. Invoke with /security-audit.

Evaluates technical design decisions and engineering judgment from a Staff Engineer perspective. Use to rubber-duck your own code or architecture, validate technical decisions before committing, or review architectural proposals and design docs. Use when user mentions "staff review", "design review", "architecture review", "technical decisions", or invokes /staff-review.

irega's coding philosophy, conventions, and preferred approaches. Always active — apply these rules for any commit, code, or PR in this session.

Generate PR description from template and git changes. Use when creating a PR and want to auto-generate description based on .github template, commits, and diffs. Extracts Jira tickets from commit messages.

Compact the current conversation into a handoff document for another agent to pick up.

Ultra-compressed communication mode. Cuts token usage ~75% by speaking like caveman while keeping full technical accuracy. Supports intensity levels: lite, full (default), ultra, wenyan-lite, wenyan-full, wenyan-ultra. Use when user says "caveman mode", "talk like caveman", "use caveman", "less tokens", "be brief", or invokes /caveman. Also auto-triggers when token efficiency is requested.