Interview stakeholders to clarify ambiguous or underspecified requirements before writing code. Use when receiving a new task, feature request, or bug report that lacks clear acceptance criteria. Produces clarified requirements ready for spec authoring. Common triggers: "new feature", "build me", "implement", "add support for", or any task where requirements are vague or incomplete.

Verify that implementation and tests remain synchronized with specs after code changes. Use when code has been generated or modified from specs, after implementation is complete, or when reviewing a PR that touches spec-covered code. Reports mismatched targets, broken test links, and undocumented behavioral changes. Common triggers: "verify the spec", "check spec alignment", "are specs up to date", or after completing implementation work.

Create or update .spec.md files from clarified requirements. Use when requirements have been gathered and confirmed, and specs need to be written or updated before implementation begins. Produces well-structured spec files with frontmatter, requirements, and test links. Common triggers: "write the spec", "update the spec", "create a spec for", or after requirement-gathering completes.

Use when the user asks to write unit tests, generate a test suite, or migrate/convert Jest or Vitest tests to Themis in Node.js/TypeScript repos. Produces Themis-native tests, runs validation commands, and applies Themis migration workflows.

Use when the user asks to run E2E tests, verify a web page, generate tests for an app, prove destructive actions are blocked, check if a UI element is visible, fill out a form, or drive a browser with natural language. Returns per-step results with safety classifications, policy decisions, DOM diffs, structured page context, and a signed audit trail.

Executes EXPLAIN ANALYZE via MCP to debug slow queries, identify missing indexes, detect sequential scans, and optimize query plans. Use when debugging slow SQL, analyzing query plans, finding unused indexes, optimizing Postgres queries, or investigating index-not-used issues.

designing a new programming language

Build OpenTelemetry collector configs, instrument services, transform telemetry with OTTL, and debug missing traces, metrics, or logs. Use for OTel/otelcol collector config, OTLP export, SDK instrumentation, sampling, cardinality, TLS/PII controls, Kubernetes/Helm values.yaml deployments, collector health and alerts, and AI coding-agent telemetry (Claude Code, Codex, Gemini CLI, GitHub Copilot).

Use when the user asks to log work, record what was done, or save task progress. Creates a structured markdown task log in `agent-logs/` with validation, waivers, and a final gates summary.

Use when the user asks you to implement a feature, fix a bug, or complete a repository task end-to-end (from scoping through code changes, validation, and final gate summary).

Use when `.tf` or `.tfvars` files have been edited, added, or removed and you need to verify that the code changes produce the intended Terraform plan before the task is considered complete. Runs `terraform plan` in `examples/test_app` and cross-checks the result against the change set you expected your edits to cause. This is a gate: do not declare a Terraform task done without passing it.

Scaffolds a complete A2A agent specifically configured for Gemini Enterprise compatibility, including the JSON-RPC root path and a health check.

Generates micro Flutter/Dart UI components (atoms/molecules) following Atomic Design patterns. Uses StatelessWidget for atoms and HookConsumerWidget for molecules, including form validation classes and auto-generated widget tests with mocktail mocks. Use when the user asks to create small Flutter UI components such as buttons, text fields, inputs, cards, or reusable widgets following Atomic Design principles, or when they need a Flutter atom, molecule, or micro-UI component with Riverpod integration.

Inspects database health by analyzing slow queries, identifying bottlenecks, and checking Postgres performance indicators. Use when diagnosing database issues, checking slow queries, inspecting DB health, finding performance bottlenecks, or troubleshooting Supabase Postgres problems.

Audits a Claude Code skill for security risks in three modes: before download (from a URL or install command), after download but before install (from a .skill file), or after install (from a local skills directory). Use this skill whenever a user is about to install a skill from any source — including GitHub URLs, git clone commands, npx/npm commands, curl/wget downloads, pip installs, marketplace links, or raw SKILL.md URLs. Also trigger when a user asks "is this skill safe?", "should I trust this skill?", "can you check this before I install it?", "audit this skill", or pastes any link to a skill repository or .skill file. If a user mentions installing ANY skill, proactively offer to audit it first — do not wait for them to ask.

Creates Postgres INSERT triggers that fire Edge Functions for document chunking and embedding generation. Configures the ingestion pipeline from raw document insert to vector storage. Use when building RAG ingestion, embed-on-insert pipelines, database-driven document ingestion, or automated embedding workflows.

Generates a Mermaid flowchart diagram of dbt model lineage using MCP tools, manifest.json, or direct code parsing as fallbacks. Use when visualizing dbt model lineage and dependencies as a Mermaid diagram in markdown format.

Use when a user needs help triaging dbt-core to dbt v2 migration errors. Runs dbt-autofix first, then classifies remaining errors into actionable categories (auto-fixable, guided fixes, needs input, blocked).

Use when migrating a dbt project from one data platform or data warehouse to another (e.g., Snowflake to Databricks, Databricks to Snowflake) using dbt Fusion's real-time compilation to identify and fix SQL dialect differences.

Build, scaffold, and demo Java AI agent systems with the langchain4j-agentic module — workflow patterns (sequential, loop, parallel, conditional), supervisor, and custom Planner strategies including the flagship typed-verdict (CriticResult-style) critic pattern, plus langchain4j-mcp tool servers and langchain4j-agentic-a2a remote agents. Use whenever the user mentions LangChain4j, langchain4j-agentic, Java AI agents, @Agent / @Tool, AgenticScope, AgenticServices, sequenceBuilder / loopBuilder / supervisorBuilder / plannerBuilder, MCP or A2A in Java, or wants to build a conference demo / workshop / POC around autonomous Java agents. Pinned to 1.15.0 core + 1.15.0-beta25 agentic/mcp; refresh with scripts/check_versions.sh.

Creates RLS policies enforcing tenant_id isolation on shared-schema tables. Verifies ALTER TABLE ENABLE ROW LEVEL SECURITY before policy creation. Requires tenant_id column and custom-access-token-hook JWT claims. Use when implementing multi-tenant data isolation, tenant-safe queries, shared schema RLS, or tenant_id policy creation.

Use when the user proposes — or the agent detects — a deviation from constitutional defaults that requires an Architecture Decision Record. Triggered by proposals to extract microservices, drop foreign keys, denormalize without measured evidence, store transactional truth in Neo4j, skip Row Level Security, skip the transactional outbox, run destructive migrations, use database-per-service, or any explicit override of a constitutional principle. Drafts a structured ADR with context, decision, consequences, alternatives rejected, migration path, validation criteria, and constitutional sections affected — and refuses to proceed with the underlying work until the ADR is at least Proposed status.

Use when designing or reviewing database architecture for high-demand multi-tenant commerce platforms — including PostgreSQL schema design, foreign keys, indexes, JSONB usage, multi-tenant isolation with Row Level Security, transactional outbox, Neo4j GraphRAG projections, event sourcing decisions, audit logging, partitioning, expand/contract migrations, and product/inventory/order modeling for restaurants, boutiques, drugstores, retailers, distributors, grocery, hardware, or appliance businesses. Triggered by any request to design tables, design schemas, create migrations, model products/variants/inventory/orders/payments, choose between monolith and microservices, choose between PostgreSQL and Neo4j as source of truth, model multi-tenant data, design event flows, or review an ER diagram.

Use when reviewing or designing how Neo4j and GraphRAG interact with PostgreSQL transactional truth — including any feature involving recommendations, semantic product search, ingredient relationships, substitution suggestions, complementary products, AI-assisted discovery, vector search combined with graph traversal, or any proposal that puts orders, inventory, payments, prices, or tenant access rules into Neo4j. Evaluates architectural proposals for data boundary violations, identifies sync pattern errors between Neo4j and PostgreSQL, produces structured design review feedback with severity-ranked findings, counter-proposals with Mermaid diagrams, eventing changes, and re-projection plans. Triggered by mentions of GraphRAG, Neo4j, knowledge graph, recommendations engine, semantic search, vector + graph hybrid search, AI product discovery, or any design that crosses the PostgreSQL ↔ Neo4j boundary.

Configures Edge Functions to use EdgeRuntime.waitUntil for background work while returning immediate 200 OK. Routes massive async workloads through pg_cron discovery and pgmq task queuing. Use when building background processing, async third-party calls, long-running edge function tasks, or webhook-driven pipelines.