Generate clean, scalable, production-ready code for a feature. Returns a structured prompt the host LLM executes to produce code, tests, logging, error handling, and docs following industry best practices.

Scaffold a new Comviva Spring Boot microservice following company standards (packaging, logging, Kafka, Consul config, error handling, OpenAPI). Returns a structured prompt that yields a complete file tree with full source for pom.xml, the application class, config beans, controllers, producers/consumers, entities, application.yml/bootstrap.yml, logback-spring.xml, Dockerfile, and tests.

Generate a complete JUnit 5 + Mockito test class for one Java class, with @DisplayName annotations on every test class, nested group, and test method for readable test reports. Follows Comviva standards (copyright header, package layout). Call once per class; for 'tests for all classes in the project', the host LLM should glob the sources and invoke this tool per file.

Review a pull request or diff. Returns a structured prompt that yields severity-tagged, file:line-anchored findings across correctness, security, performance, maintainability, and tests.

Analyze symptoms, stack traces, logs, and source to find the root cause of a bug. Returns a structured prompt that yields a hypothesis with evidence, severity, priority, recommended fix, and prevention actions.

Generate a detailed Root Cause Analysis report for a production incident. Returns a structured prompt the host LLM completes into a publish-ready RCA document with summary, timeline, root cause, impact, resolution, corrective actions, preventive measures, and lessons learned.

Fetch a Jira issue by key.

Search Jira issues with a JQL expression.

Create a Jira issue. Useful for filing bugs after analysis, opening tracking tickets for RCA action items, or creating stories from implementation plans.

Update a Jira issue's fields.

Transition a Jira issue to a new status (e.g. 'In Progress', 'Done'). Optionally posts a comment with the transition.

Add a comment to a Jira issue. Use to attach investigation notes, RCA summaries, or implementation updates.

Link two Jira issues. Link types are project-defined; common ones: 'Relates', 'Blocks', 'is blocked by', 'Caused by', 'Duplicate'.

Attach a remote link (e.g. PR URL, deployment URL, dashboard URL) to a Jira issue.

Recommend a scalable system design for a stated goal. Returns a structured prompt covering architecture decisions, component diagram-as-text, data flow, scaling strategy, failure modes, and an explicit decision log.

Review an existing architecture (API, database, microservice, event pipeline, or cloud infra) for scalability, reliability, security, and operability. Returns a structured prompt that yields graded findings and prioritized remediation.

Identify technical debt in a codebase and propose a refactoring strategy. Returns a structured prompt covering debt inventory, business impact, refactor roadmap, and quick wins.

Validate a CI/CD pipeline configuration for correctness, security, efficiency, and best practices. Returns a structured prompt yielding severity-tagged findings and recommended fixes.

Review Kubernetes, Docker, Terraform, Helm, or CloudFormation configuration for correctness, security, reliability, and cost. Returns a structured prompt with anchored findings and remediation.

Perform a pre-deployment readiness check. Returns a structured prompt assessing risk, validating rollback, observability, and go/no-go gates.

Generate technical documentation — design doc, API reference, runbook, README, or module overview — from supplied source material. Returns a structured prompt the host LLM completes into the document.

Generate an Architecture Decision Record (ADR) in the MADR-style format. Returns a structured prompt the host LLM completes into a fully-formed ADR.

Generate a phased implementation plan for a feature or change. Returns a structured prompt yielding milestones, dependencies, risks, and a definition of done.

Generate a migration strategy from one technology / version / pattern to another. Returns a structured prompt yielding stages, dual-write/dual-read plans, validation, cutover, and rollback.