hades-openspec-java-workflow

Use when a repository has CodeGraph initialized or the task needs accurate codebase navigation, impact analysis, call-chain tracing, symbol lookup, architecture exploration, or prevention of AI mis-locating Java/Spring/MyBatis code. Prefer CodeGraph context before grep/read loops when .codegraph exists.

Use when CodeGraph CLI, MCP wiring, or project indexing may be missing, before using CodeGraph for code navigation, impact analysis, call chains, symbol lookup, or when helping install or initialize CodeGraph on macOS, Linux, or Windows.

Use before writing, modifying, refactoring, or reviewing Java/Spring/MyBatis code in Hades Openspec Java Workflow, especially when scope could drift, requirements are ambiguous, implementation may overgrow, or verification evidence is needed.

Use when a request involves internal database MCP access, schema inspection, business data lookup, SQL execution, validation queries, database write approval, DELETE/DDL safety, or database evidence for PRD/OpenSpec implementation.

Use when implementing tasks from an OpenSpec change, especially /opsx:apply or natural language requests such as execute plan, start implementation, or apply this change.

Use when a request may come from Asana, Notion, user conversation, Figma, Stripe, Cloudflare, GitHub PRs, or internal context and should be routed through PRD, OpenSpec, implementation, review, or archive. Also use when deciding whether to read external context.

Use when reviewing Java, Spring Boot, MyBatis, SQL, transaction logic, backend API changes, or PRs before merge. Focus on correctness, transaction semantics, null safety, concurrency, security, logging, API performance, SQL quality, and maintainable layered design.

Use when Java, Maven, Gradle, Spring Boot build, compilation, dependency, annotation processor, Checkstyle, SpotBugs, PMD, or test compilation fails. Fix build errors with minimal surgical changes and rerun verification after each fix.

Use when writing or reviewing Java/Spring/MyBatis code in Hades Openspec Java Workflow, including naming, package structure, DTO/VO/PO boundaries, null safety, exceptions, logging, collections, MyBatis SQL, and tests.

Use when adding, changing, reviewing, or preparing PRs for Java/Spring/MyBatis code. Decides required unit tests, integration tests, regression tests, manual verification, test commands, and PR test evidence.

Use when a request is a large refactor, architecture change, module rewrite, cross-module migration, technical debt cleanup, framework upgrade, API compatibility migration, or any change too large for a single OpenSpec change. Forces Discovery, RFC, characterization tests, phased OpenSpec changes, small PRs, compatibility, rollback, and cleanup.

Use when a task involves MySQL MCP access, AI database accounts, SQL execution, table/schema inspection, data writes, DELETE authorization, migration SQL, MyBatis SQL validation, EXPLAIN review, SQL performance evidence, or database safety gates for Java/Spring/MyBatis work.

Use when the user asks to generate, continue, inspect, apply, execute, validate, review, archive, or otherwise act on an OpenSpec change using natural language or /opsx commands.

Use when OpenSpec CLI may be missing or uninitialized, before running openspec or /opsx commands, or when helping install or initialize Fission AI OpenSpec on macOS, Linux, or Windows.

Use when proposing or creating an OpenSpec change and generating artifacts from confirmed PRD documents, especially for /opsx:propose or user requests like generate OpenSpec or create plan.

Use after OpenSpec implementation code changes are complete and before final verification or review, especially when checking code against PRD/OpenSpec/tasks, scope drift, documentation sync, Java/Spring/MyBatis/MySQL quality gates, tests, EXPLAIN evidence, or rollback readiness.

Use before opening or merging a PR for a Java/Spring/MyBatis requirement delivered through Hades PRD and OpenSpec. Checks post-implementation self-check, build, tests, OpenSpec tasks, PR description, review readiness, API/SQL performance evidence, risk, rollback, and documentation sync.

Use when deciding which PRD split documents to create for a feature based on affected areas such as UI behavior, API contracts, database/schema impact, or backend technical design.

Use when turning a requirement, Asana task, Notion note, user conversation, product idea, bug, incident, design request, or technical change into Chinese PRD documentation before OpenSpec or implementation.

Use after implementing, post-implementation self-checking, and verifying an OpenSpec change, when prompting the user for review, running code review, handling review feedback, or deciding whether to revise, continue, or archive.

Use when adding or reviewing Spring Boot authentication, authorization, user input handling, API endpoints, file uploads, imports/exports, secrets, CORS/CSRF, rate limiting, sensitive data, SQL injection risk, or dependency security.

Use when designing or implementing Spring Boot backend services, REST APIs, controller-service-mapper layering, validation, exception handling, transactions, pagination, caching, async jobs, external calls, and observability in Hades.

Use whenever designing, writing, modifying, or reviewing Java/Spring/MyBatis SQL, Mapper methods, list/search/statistics/export APIs, pagination, sorting, dynamic query conditions, indexes, or data-access loops. In implementation mode, follow it as coding rules before writing SQL; in review mode, use it to review code; when the user explicitly invokes this skill/prompt, do a SQL-only review and do not expand into full PRD/OpenSpec/Java review or PR Gate unless asked.

Use when requirements, workflow rules, scope, non-goals, acceptance criteria, API contracts, database rules, state machines, implementation constraints, security rules, review rules, verification rules, PRD docs, or OpenSpec artifacts change during a conversation.