CI/CD pipelines, Docker, Kubernetes, cloud infrastructure, monitoring, deployment automation. Use for GitHub Actions, containerization, IaC, secrets management, observability.
Debug systematically with root cause analysis before fixes. Use for bugs, test failures, unexpected behavior, performance issues, CI/CD failures, database diagnostics.
ALWAYS activate before fixing ANY bug, error, test failure, CI/CD issue, type error, lint, log error, UI issue.
Analyze codebase and manage project documentation — init, update, summarize. Use for creating or updating docs/ directory.
Semantic code analysis with GitLab Knowledge Graph. Use for go-to-definition, find-usages, impact analysis, architecture visualization. Supports Ruby, Java, Kotlin, Python, TypeScript/JavaScript.
Git operations with conventional commits. Use for staging, committing, pushing, PRs, merges. Auto-splits commits by type/scope. Security scans for secrets.
Build React/TypeScript frontends with modern patterns. Use for components, Suspense, lazy loading, useSuspenseQuery, MUI v7 styling, TanStack Router, performance optimization.
Create polished frontend interfaces from designs, screenshots, or videos. Use for UI replication, 3D experiences, prototypes, immersive interfaces, avoiding AI slop.
Execute an approved implementation plan exactly and safely. Use when a plan already exists (for example in docs/plans/...) and work must be carried out phase-by-phase with verification checkpoints, status tracking, and final execution reporting.
Initialize or update project documentation. Use when setting up docs for a new project (`--init`) or syncing docs to current codebase and architecture changes (`--update`).
Rapid implementation for small, low-risk, well-defined changes. Use when the task is narrow in scope, has clear acceptance criteria, and can be completed safely without a formal multi-phase plan.
Create a local %erg v1 file for agent coordination.
Claim a local ticket for work (cross-worktree safe).
List local tickets that are ready for work (unblocked, unclaimed).
Repo healthcheck — git hygiene, test status, and deep freshness verification of status/directive docs. Gracefully degrades when project-specific conventions (git-erg tickets, STATE.md, etc.) are absent.
Project rules and conventions for the Imperial Dragon workflow. Auto-invoked at session start to establish behavioral constraints: worktree isolation, git discipline, coding standards, and STATE.md format.
End-of-day session wrap-up. Pushes branches, runs tests, refreshes STATE, offers autonomous session.
Write, update, or sweep persistent memory. Enforces list caps, TTLs, and staleness criteria.
Re-resolve every DOI/URL/eprint in a related-work-note's Bibliography. Append a provenance line to Methods. One-line verdict to stdout (PASS / WARN / FAIL).
Run Imperial Dragon batch across multiple tickets. Picks targets, manages waves, enforces isolation. Always autonomous — never merges.
Begin work on a ticket. Creates worktree, writes first test, transitions to Execute phase.
Terse caveman mode. Levels: lite, full, ultra. Use when user says "use caveman", or invokes /caveman. Auto-triggers when token efficiency is requested.
Analyze images/audio/video with Gemini API. Generate images (Imagen 4), videos (Veo 3). Use for vision analysis, transcription, OCR, design extraction, multimodal AI.
Bootstrap new projects with research, tech stack selection, design, and scaffolding. Modes: full (interactive), auto (default), fast (skip research).
Add authentication with Better Auth (TypeScript). Use for email/password, OAuth providers, 2FA/MFA, passkeys/WebAuthn, sessions, RBAC, rate limiting.
Build backend APIs, services, and infrastructure. Use for REST/GraphQL APIs, database models, authentication, middleware, background jobs, caching.
Answer technical and architectural questions with expert consultation. Use for system design, technology decisions, architectural guidance.
ALWAYS activate before implementing ANY feature, plan, or fix. End-to-end implementation with automatic workflow detection.
Brainstorm solutions with trade-off analysis and brutal honesty. Use for ideation, architecture decisions, technical debates, feature exploration, feasibility assessment.
Search library/framework documentation via llms.txt (context7.com). Use for API docs, GitHub repository analysis, latest library features, technical documentation lookup.