kumaran-is/claude-code-onboarding
Agent Claude Code
RAG pipeline code reviewer. Use when reviewing retrieval pipelines, embedding generation code, chunking strategies, or semantic search implementations. Checks: embedding model pinned, chunking strategy justified, hybrid vs pure-vector choice documented, reranking present for production, null guards on retrieval, and…
kumaran-is/claude-code-onboarding
Agent Claude Code
Final validation gate for UI and feature work. Defaults to NEEDS WORK — APPROVED requires visual screenshot evidence and passing tests, not assertions. Use before merging any PR that touches UI, user flows, or API behaviour visible to end users. Fantasy-allergic, evidence-obsessed, binary verdict only.
kumaran-is/claude-code-onboarding
Agent Claude Code
Mine Reddit threads to extract real customer pain points, unmet needs, and product gaps. Outputs a ranked top-10 opportunities report. Use when user says: find pain points for [topic], reddit research [topic], customer complaints about [topic].
kumaran-is/claude-code-onboarding
Agent Claude Code
Agent that performs Riverpod pattern verification for Flutter applications. Specializes in provider types, ref usage, AsyncValue handling, and lifecycle analysis. Examples:\n\n \nContext: Several new Riverpod providers and AsyncNotifiers were added to a Flutter profile feature.\nUser: "Review the Riverpod providers I…
kumaran-is/claude-code-onboarding
Agent Claude Code
Scans .claude/skills/ for behavioral patterns appearing in 2+ skills that are actionable, violation-prone, and not already in .claude/rules/. Presents candidates for promotion with evidence and verdict options. Use when you want to keep rules files up-to-date as the skill library grows.
kumaran-is/claude-code-onboarding
Agent Claude Code
Security vulnerability detection and remediation specialist. Flags secrets, SSRF, injection, unsafe crypto, and OWASP Top 10 vulnerabilities. Examples:\n\n \nContext: New authentication endpoints and input handling code were just implemented.\nUser: "Check the new auth endpoints for security…
kumaran-is/claude-code-onboarding
Agent Claude Code
Use when reviewing code that contains error handling, catch blocks, fallback logic, or any code that could suppress errors. Proactively invoke after implementing features with API calls, database operations, external service integrations, or any I/O that can fail. Enforces code-standards.md "No Silent Failures" rule…
kumaran-is/claude-code-onboarding
Agent Claude Code
Reviews a newly created or modified Claude Code skill against the writing-skills spec. Checks Iron Law presence, description quality (CSO trigger words), allowed-tools declaration, body line count (≤500), progressive disclosure structure, forbidden files, and cross-reference patterns. Use after creating or modifying…
kumaran-is/claude-code-onboarding
Agent Claude Code
Specialized code reviewer for Java 21 Spring Boot 3.5.x WebFlux reactive services. Reviews for reactive correctness, Resilience4j patterns, security, testing, and production readiness. Examples:\n\n \nContext: New reactive service and controller were written for user management in a Spring Boot WebFlux project.\nUser…
kumaran-is/claude-code-onboarding
Agent Claude Code
Red-Green-Refactor enforcement agent. Use when implementing new features with test-first methodology, verifying 80%+ coverage gates, or auditing existing tests for edge case completeness. Covers Spring Boot/JUnit5, Python/pytest, NestJS/Vitest, and Flutter/fluttertest.
kumaran-is/claude-code-onboarding
Agent Claude Code
Expert Terraform specialist for Google Cloud Platform infrastructure. Provisions and manages Cloud Run services, Cloud SQL (PostgreSQL), Artifact Registry, Firebase projects, IAM bindings, Secret Manager, and Workload Identity Federation. Masters GCS remote state, multi-environment (staging/prod) module design…
kumaran-is/claude-code-onboarding
Agent Claude Code
Expert in threat modeling methodologies, security architecture review, and risk assessment. Masters STRIDE, PASTA, attack trees, and security requirement extraction. Use when designing new systems, reviewing architecture for security gaps, preparing for security audits, or identifying attack vectors before…