← ClaudeAtlas

backend-code-review-workflowlisted

Orchestrate a consolidated backend review across Kotlin/Spring correctness, refactoring, design patterns, object-oriented design, and security. Use for pull requests, diffs, or backend modules that need one severity-gated report without losing specialist dissent.
ch4570/vulpora · ★ 1 · Code & Development · score 67
Install: claude install-skill ch4570/vulpora
# Backend Code Review Workflow Run the mandatory reviewers against one immutable scope, then reconcile their evidence. Review only; do not modify code unless the user separately requests implementation. ## Exact dependencies - Skills: `kotlin-spring-review`, `refactoring-catalog`, `design-pattern-apply`, `oop-design-review` - Agent: `security-auditor` Treat every dependency as mandatory. Do not substitute a similarly named skill or agent. ## Cost-aware native routing Resolve portable profiles from the current runtime capability catalog at dispatch. Use `kotlin-spring-review=standard/medium`, `security-auditor=standard/medium`, and `refactoring-catalog|oop-design-review|design-pattern-apply=frugal/low`. Apply a `frontier/high` floor only to the affected security pass for a confirmed authn/authz boundary, irreversible data, or public-contract risk. The workflow owner and reconciliation default to `standard/medium`; the number of installed rules or skills is not a reason to require a frontier primary. A smaller route remains safe only when its scope and evidence packet are bounded as described below. Use `frontier/high` for the affected judgment, not for every pass or for orchestration as a whole. Every native handoff uses `model_selection: explicit-native-override`; spawn with `fork_turns: none`, the exact resolved `model`, and exact `reasoning_effort`. Never inherit the primary model or reasoning setting. Record requested/observed route fields in the run ledger and m