Plugin Claude Code
Senior Dev Kit — a senior engineering team for Claude Code.
Plugin Claude Code
Senior Dev Kit — a senior engineering team for Claude Code.
Plugin Claude Code
Guard agents, auto-triggered skills and path-scoped rules that make Claude Code behave like a senior engineering team.
Settings file Claude Code
Agent settings configuring $schema, permissions.
Instructions file
Instructions for mtvrkan/senior-dev-kit, covering senior-dev-kit — repo claude.md, commands, conventions, effort strategy and context budget (this dev's environment only).
Agent
Highest-priority signal wins. Read top-to-bottom; stop at the first match.
Agent
Use for localized bugs, runtime errors, failing tests, console errors, regressions, or broken behavior where root cause can be isolated. Escalate protected areas.
Agent
Use for database schema changes, data modeling, ORM queries, indexes, constraints, transactions, and migration deployment safety (destructive changes, rollback, backward compatibility, production risk, deployment order, backups). Read-only — produces a plan, waits for approval; never executes migrations.
Agent
Use when a project's visual design has to be decided rather than matched — a first page or screen with no DESIGN-SPEC.md, a redesign, a brief with references or brand assets, or "make it look like its own product". Produces the direction, the tokens and the signature moment; hands construction to ui-fixer.
Agent
CI/CD, Docker, Terraform, Kubernetes, and infrastructure change guardian — plans first, requires explicit approval for destructive or production-affecting changes. Use for: GitHub Actions, GitLab CI, K8s manifests, Helm charts, docker-compose, deployment scripts.
Agent
Use for slow queries, N+1, bundle size, caching, render loops, memory leaks, latency, and expensive computations.
Agent
Use for auth, authorization, payment, billing, input validation, secrets, injection risks, session/token handling, rate limiting, sensitive data, and security reviews — plus tool-driven dependency, secret, SAST, and container/filesystem scans. Read-only by default — delegates implementation to senior-engineer.
Agent
Use for scoped medium feature implementation or safe refactors requiring multiple files, tests, and existing project patterns. Do not use for critical protected changes without a plan.
Agent
Use for low-risk frontend-only UI changes — modals, buttons, layout, responsive styling, Tailwind/CSS, component polish, new pages, new screens. Do not use for backend, auth, payment, database, migrations, secrets, or CI.
Command
Audit built UI against WCAG 2.2 AA — keyboard, focus, contrast, state announcement, targets, reflow.
Command
List all installed Senior Dev Kit agents and when to use each one.
Command
Audit architecture integrity — boundary violations, dependency direction, mixed patterns, contract drift.
Command
Audit built UI for design-direction adherence, generic-output tells, and layout monotony.
Command
Audit the project for SEO, AEO, Core Web Vitals, and technical SEO issues.
Command
List all installed Senior Dev Kit skills and when each one auto-triggers.
Hook
Runs when a session starts, executing session-context.mjs via node. From mtvrkan/senior-dev-kit.
Skill Claude CodeCodex
Use for designing REST/GraphQL API contracts before implementing, and for planning breaking-change version transitions (v1→v2, deprecations). No code edits.
Skill Claude CodeCodex
Use for localized bugs, runtime errors, white screens, failing tests, console errors, and regressions. Make the smallest safe fix.
Skill Claude CodeCodex
Use to review diffs without editing code. Focus on bugs, regressions, security risks, data loss, missing validation, missing tests, and performance risks.
Skill Claude CodeCodex
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.