mtvrkan/senior-dev-kit

codebase-overview

Use to generate or refresh a codebase architecture overview — directory map, data flow, key component responsibilities — so future research reuses it instead of rediscovering the codebase.

feature-plan

Use before implementing large, multi-file, architecture, DB, auth, payment, or unclear features. Produces a detailed plan; no code.

from-scratch

Use when starting a new project from scratch. Establishes contracts, design system, and architecture skeleton before any feature code. Enforces phase gates to prevent incoherence and bugs.

kit-setup

Install the parts of Senior Dev Kit a plugin cannot carry — the path-scoped rules and the deny list — into the user's Claude Code settings directory.

new-page

Use when building a brand-new admin panel page or screen from scratch in ANY web framework. Enforces design quality gates. Do NOT use for modifying existing pages — use ui-change instead.

new-screen

Use when building a brand-new mobile screen, bottom sheet, or major UI section from scratch in Kotlin/Compose, Flutter, or Swift/SwiftUI. Enforces design quality gates. Do NOT use for modifying existing screens — use ui-change instead.

refactor-safe

Use for safe refactors that preserve behavior. Keep diffs small and verification strong.

ui-change

Use for small UI changes: modal, button, responsive layout, component styling, Tailwind/CSS. Do not use for backend, auth, database, or payment.

api-design

Use for designing REST/GraphQL API contracts before implementing, and for planning breaking-change version transitions (v1→v2, deprecations). No code edits.

bug-fix

Use for localized bugs, runtime errors, white screens, failing tests, console errors, and regressions. Make the smallest safe fix.

code-review

Use to review diffs without editing code. Focus on bugs, regressions, security risks, data loss, missing validation, missing tests, and performance risks.

db-change

Use for database schema/model/query changes AND new entity/relation/document modeling — SQL, NoSQL, ORM, indexes, constraints, data safety. Once a migration file exists, hand off to migration-review.

deep-research

Deep multi-source research — fan-out searches, fetch primary content, cross-verify claims, synthesize a cited report. Invoke via /deep-research.

docs-update

Use for README, setup instructions, changelog, API docs, and short documentation updates.

env-audit

Audits environment variables across the codebase — missing declarations, leaked defaults, unused vars, .env.example drift. Invoke via /env-audit.

feature-build

Use for scoped medium feature implementation after task is clear. Prefer one small diff and existing project patterns.

incident-response

Use to coordinate a live production incident or outage — triage severity, decide which guards to dispatch in what order, keep one timeline. Produces the dispatch plan only; the caller invokes each guard.

kit-doctor

Use to diagnose a senior-dev-kit installation — component file counts, settings merge state, and version drift. Read-only.

migration-review

Use for migration files, destructive changes, production data risk, rollback strategy, and backward compatibility. Deployment-safety phase that follows db-change once a migration file exists.

performance-check

Use for slow code, slow queries, bundle size, caching, N+1, render loops, memory, and latency issues.

project-memory

Use to persist durable project-level facts (decisions, gotchas, non-obvious conventions) to a project memory file so future sessions don't rediscover them. No code changes.

release-gate

Use before release/deploy to run a GO/NO-GO safety checklist — build, tests, migrations, env vars, security scan, backward compat, rollback, known risks.

security-review

Use for auth, permissions, payment, secrets, input validation, session/token handling, SQL/NoSQL injection, and sensitive user data.

security-scan

Use for dependency audits (CVEs, outdated majors, paid licenses, bloat), secret scans, SAST, and container/filesystem scans. Not for low-risk UI-only tasks.

test-writer

Use to add or update tests using the existing test framework for changed behavior, edge cases, regressions, or existing untested code the user wants covered.