Backend execution planning skill used after PRD or bug context exists and before coding. Produces an implementation-ready, risk-aware, validation-ready backend plan with explicit checkpoints.
Review backend-ready patches for correctness, security, maintainability, and testing risk, then provide evidence-based findings with severity and delivery recommendation.
Root-level lightweight verification for docs/config/rules/templates/AGENTS/skills and small-scope low-risk non-invasive changes; confirms completeness, consistency, references, paths, command claims, and unresolved risks with explicit VERIFIED/UNVERIFIED/GAP status.
Use when entering an unfamiliar repo or before generating/updating AGENTS.md and skills; scan the full repository (monorepo-safe), map real structure/commands/constraints without guessing, and produce a reusable architecture snapshot for future agent workflows.
Use this skill when backend requirements are ambiguous before implementation. Convert user intent into a structured, testable backend PRD with explicit scope, non-goals, API/data/auth impact, assumptions, risks, and open questions.
Backend refactoring workflow for improving structure, readability, maintainability, duplication, module boundaries, and testability without changing externally observable behavior. Use when scope is bounded and behavior-preservation can be verified; do not use for feature implementation, API/schema/security redesign, dependency upgrades, or broad system redesign.
Implement approved backend code changes in existing Spring Boot modules with minimal safe edits. Use only after task scope and acceptance criteria are clear. Reuse current controller/service/repository/dto/mapper/config patterns, preserve API contracts unless explicitly approved, run relevant backend verification commands, and report evidence, risks, and unresolved items.
Backend scan/architecture discovery skill for the backend/ subtree. Use before backend implementation/review/refactor to build a reusable, evidence-based architecture snapshot (facts, inferences, unknowns) without changing production behavior.