potenlab-workflow

Generate progress.json from finalized plans

Post-completion adjustments tracked in changes.json

Execute all tasks in a phase with wave-based parallel coder agents (max 4 per wave)

Generate Vitest test files from test-plan.md into root tests/ directory

Confirm potenlab-workflow is installed and active

Display Potenlab workflow overview and status

Orchestrate project planning with specialist agents (includes test plan generation)

Edit existing plans based on user feedback

Run all Vitest tests and generate test.result.json

Run Vitest tests for a specific phase

Resolve pending security findings from security-list.json

Run a comprehensive security audit on the project and generate security-list.json with findings

Detect code changes and update tests to stay in sync

Creates backend-plan.md by reading dev-plan.md and translating its backend tasks into detailed schema definitions, migration SQL, RLS policies, query specifications, and connection strategies. Uses MCP Postgres to discover current database state and enforces Supabase best practices. Does NOT write code — coder agents execute from this plan. Examples: <example> Context: User has a dev-plan.md and needs backend planning. user: "Create the backend plan from the dev plan" assistant: "I'll use the backend-specialist agent to read dev-plan.md and create backend-plan.md with schema specs and RLS policies." <commentary> Since the user needs a backend plan, use the backend-specialist to produce backend-plan.md. </commentary> </example> <example> Context: User wants to plan database schema. user: "Plan the database tables from the dev plan" assistant: "I'll use the backend-specialist agent to create backend-plan.md with table definitions, indexes, and RLS policies." <commentary> Since the user needs schema planning, use the backend-specialist with MCP Postgres tools. </commentary> </example> <example> Context: User wants to review their schema design. user: "Review my database schema against best practices" assistant: "I'll use the backend-specialist agent to audit against Supabase Postgres best practices and update backend-plan.md." <commentary> The user wants best practice validation. Use the backend-specialist which has the full reference. </commentary> </example>

Creates frontend-plan.md by reading dev-plan.md and translating its frontend tasks into detailed component specifications, file paths, props, patterns, and implementation guidelines. Uses shadcn MCP to discover components and enforces Vercel React best practices. Does NOT write code — coder agents execute from this plan. Examples: <example> Context: User has a dev-plan.md and needs frontend planning. user: "Create the frontend plan from the dev plan" assistant: "I'll use the frontend-specialist agent to read dev-plan.md and create frontend-plan.md with component specs and implementation details." <commentary> Since the user needs a frontend plan, use the frontend-specialist to produce frontend-plan.md. </commentary> </example> <example> Context: User wants to plan a specific feature's frontend. user: "Plan the frontend for the auth feature" assistant: "I'll use the frontend-specialist agent to create frontend-plan.md with detailed component specs for auth." <commentary> Since the user wants frontend planning for a feature, use the frontend-specialist. </commentary> </example> <example> Context: User wants to review frontend architecture. user: "Review my frontend architecture against best practices" assistant: "I'll use the frontend-specialist agent to audit against Vercel React best practices and update frontend-plan.md." <commentary> Since the user wants best practice review, use the frontend-specialist which has the full reference. </commentary> </example>

Executes complex, multi-file tasks from dev-plan.md. Uses Opus for deep reasoning on tasks requiring cross-file coordination, architectural awareness, and intricate business logic. Handles the heavy lifting that small-coder cannot. Only writes code — does NOT update progress.json. Examples: <example> Context: User wants a full feature module implemented. user: "Build the auth feature from task 3.1" assistant: "I'll use the high-coder agent to implement this multi-file feature module." <commentary> A full feature with api/, components/, hooks/, and types/ is a complex task — use high-coder. </commentary> </example> <example> Context: User wants a page with data fetching and state management. user: "Implement the dashboard page with all its data connections" assistant: "I'll use the high-coder agent to build the dashboard with API integration and state." <commentary> A page with data fetching, multiple components, and state is complex — use high-coder. </commentary> </example> <example> Context: User wants a single utility function. user: "Add a formatDate helper" assistant: "This is a small task — I'll use the small-coder agent instead." <commentary> Single-file utility functions are NOT complex tasks — use small-coder for this. </commentary> </example>

Creates progress.json by reading dev-plan.md, frontend-plan.md, and backend-plan.md. Parses all phases and tasks into a structured JSON tracking file that coder agents use to identify pending work and dependencies. Runs AFTER all plans are finalized. Examples: <example> Context: User has finalized dev-plan.md and specialist plans. user: "Create the progress tracker from the plans" assistant: "I'll use the progress-creator agent to read all plans and generate progress.json." <commentary> Since all plans exist and the user needs progress tracking, use progress-creator to generate progress.json. </commentary> </example> <example> Context: User just finished the planning phase. user: "Generate progress.json" assistant: "I'll use the progress-creator agent to parse dev-plan.md and create progress.json with all tasks." <commentary> Since the user explicitly wants progress.json, use progress-creator. </commentary> </example> <example> Context: dev-plan.md was updated and progress.json needs to sync. user: "Sync the progress file with the updated dev plan" assistant: "I'll use the progress-creator agent to re-read dev-plan.md and regenerate progress.json." <commentary> Since dev-plan.md changed, use progress-creator to regenerate a fresh progress.json. </commentary> </example>

Generates feature tests for a Next.js + Supabase app using Vitest. Reads test-plan.md and produces .test.ts files that verify behavior via database operations and RLS policies against a running Supabase local instance. Does NOT test UI, does NOT use Playwright, does NOT create or alter tables. Examples: <example> Context: User has a test-plan.md and wants tests generated. user: "Generate tests from the test plan" assistant: "I'll use the qa-specialist agent to read test-plan.md and generate Vitest test files for each feature." <commentary> Since the user needs tests generated from a plan, use the qa-specialist to produce .test.ts files. </commentary> </example> <example> Context: User wants to test RLS policies for a specific table. user: "Write tests for the orders table RLS policies" assistant: "I'll use the qa-specialist agent to generate RLS behavior tests for the orders table using Supabase client." <commentary> RLS policy testing via database behavior is the qa-specialist's core strength. </commentary> </example> <example> Context: User wants to test a feature's CRUD operations. user: "Test the project feature end-to-end" assistant: "I'll use the qa-specialist agent to generate CRUD behavior tests against Supabase local for the project feature." <commentary> Feature behavior tests via database operations — use qa-specialist. </commentary> </example> <example> Context: User wants UI component tests. user: "Test the dashboard page rendering" assistant: "That's a UI test — I can't help with that. Use a frontend testing approach with @testing-library/react instead." <commentary> UI testing is explicitly outside qa-specialist scope. Decline and suggest alternatives. </commentary> </example>

Performs comprehensive security audits on repositories by scanning for vulnerabilities across OWASP Top 10 categories. Reads source code, configuration files, environment setup, Supabase RLS policies, API routes, and dependencies to produce security-report.md with findings, severity ratings, and remediation steps. Specialized in Next.js + Supabase + Solana stacks. Does NOT fix code — provides actionable findings for coder agents. Examples: <example> Context: User wants a full security audit of their project. user: "Run a security audit on my project" assistant: "I'll use the security-specialist agent to scan the codebase for vulnerabilities and produce security-report.md." <commentary> Since the user needs a comprehensive security review, use the security-specialist to audit the full codebase. </commentary> </example> <example> Context: User wants to check RLS policies before deployment. user: "Audit my Supabase RLS policies for security issues" assistant: "I'll use the security-specialist agent to review all RLS policies against best practices and check for data leaks." <commentary> RLS policy auditing is a core strength of the security-specialist — use it for Supabase security reviews. </commentary> </example> <example> Context: User wants to review API routes for vulnerabilities. user: "Check my API routes for security vulnerabilities" assistant: "I'll use the security-specialist agent to scan API routes for injection, auth bypass, rate limiting gaps, and CSRF issues." <commentary> API route security scanning is within scope — use the security-specialist. </commentary> </example> <example> Context: User wants to check for hardcoded secrets. user: "Scan for any hardcoded secrets or leaked credentials" assistant: "I'll use the security-specialist agent to scan the codebase for hardcoded API keys, tokens, passwords, and secret patterns." <commentary> Secret scanning is a core capability — use the security-specialist. </commentary> </example> <example> Context: User wants pre-deployment security sign-off. user: "Is my project secure enough to deploy to production?" assistant: "I'll use the security-specialist agent to run a pre-deployment security checklist and produce a go/no-go report." <commentary> Pre-deployment security validation — use the security-specialist for the full checklist. </commentary> </example>

Executes small, isolated tasks from dev-plan.md. Uses Sonnet for fast, cost-efficient implementation of single-file changes, minor fixes, and simple feature additions. Refuses large or multi-file tasks — escalates them back to the user. Only writes code — does NOT update progress.json. Examples: <example> Context: User wants a small utility function added. user: "Add the formatDate utility from task 0.3" assistant: "I'll use the small-coder agent to implement this small utility task." <commentary> A single utility function is a small task — perfect for small-coder. </commentary> </example> <example> Context: User wants a minor component tweak. user: "Update the button variant colors from task 2.4" assistant: "I'll use the small-coder agent to update the button styles." <commentary> A single-file style change is a small task — use small-coder. </commentary> </example> <example> Context: User wants a large feature implemented. user: "Build the entire auth feature" assistant: "This is a large task spanning multiple files and concerns. I'll use the high-coder or frontend-specialist instead." <commentary> Multi-file features are NOT small tasks — do NOT use small-coder for this. </commentary> </example>

Creates dev-plan.md as the single source of truth by reading ui-ux-plan.md from the ui-ux-specialist. Translates design decisions, wireframes, component specs, and accessibility requirements into a minimal, phased development checklist with verifiable tasks. Does NOT create progress.json — that is handled by a separate agent. Examples: <example> Context: User has a ui-ux-plan.md and needs a dev plan. user: "Create a dev plan from my UI/UX plan" assistant: "I'll use the tech-lead-specialist agent to read ui-ux-plan.md and create dev-plan.md as the single source of truth." <commentary> Since the user has a ui-ux-plan.md, use the tech-lead-specialist to translate it into an actionable dev-plan.md. </commentary> </example> <example> Context: UI/UX specialist just finished the design plan. user: "The design plan is ready, now create the development plan" assistant: "I'll use the tech-lead-specialist agent to read ui-ux-plan.md and produce dev-plan.md." <commentary> Since ui-ux-plan.md exists, use the tech-lead-specialist to create the dev plan from it. </commentary> </example> <example> Context: User wants to start building from the design. user: "Turn the design plan into developer tasks" assistant: "I'll use the tech-lead-specialist agent to convert ui-ux-plan.md into a phased dev-plan.md." <commentary> Since the user wants developer tasks from the design, use the tech-lead-specialist. </commentary> </example>

Creates ui-ux-plan.md with comprehensive design strategy, user research, wireframes, and design system documentation. Ensures WCAG 2.1 AA/AAA accessibility compliance and beautiful, intuitive interfaces. Examples: <example> Context: User has a PRD file and needs a UI/UX design plan. user: "Create a UI/UX plan from my PRD" assistant: "I'll use the ui-ux-specialist agent to analyze your PRD and create ui-ux-plan.md with user research, wireframes, and design system." <commentary> Since the user has a PRD and needs UI/UX planning, use the ui-ux-specialist agent to produce a comprehensive design plan. </commentary> </example> <example> Context: User wants to improve their app's user experience. user: "Help me design a better user flow for the checkout process" assistant: "I'll use the ui-ux-specialist agent to create user journey maps, wireframes, and an optimized checkout flow." <commentary> Since the user needs UX improvements, use the ui-ux-specialist agent to analyze and design better user flows. </commentary> </example> <example> Context: User needs a design system for their application. user: "Create a design system for my SaaS dashboard" assistant: "I'll use the ui-ux-specialist agent to create a comprehensive design system with components, tokens, and accessibility guidelines." <commentary> Since the user needs a design system, use the ui-ux-specialist agent which specializes in component libraries and design tokens. </commentary> </example>

Finalizes the planning phase by spawning progress-creator agent to read all plan files (dev-plan.md, frontend-plan.md, backend-plan.md) and generate progress.json — the structured task tracker used by coder agents. Run this AFTER /plan has completed. Triggers on: complete plan, finalize plan, create progress, generate progress, progress json.

Post-completion development skill for adjustments and changes after all phases are done. Accepts change requests from arguments (/developer fix the login button) or via AskUserQuestion. Analyzes existing code and plans, creates change tasks in docs/changes.json, spawns small-coder and high-coder agents in parallel, and tracks all results in changes.json. Triggers on: developer, dev, adjust, change, fix, tweak, modify, post-dev.

Executes all tasks in a specific phase using wave-based parallel execution (max 4 agents per wave). Spawns small-coder and high-coder agents based on progress.json complexity. Updates progress.json after EACH wave so completed tasks immediately unblock dependents. Accepts phase number as argument (/execute-phase 3) or asks via AskUserQuestion. Triggers on: execute phase, run phase, build phase, develop phase, start phase.

Generates Vitest test files from test-plan.md using the qa-specialist agent. Reads the test plan and vitest best practices, then spawns qa-specialist agents to produce .test.ts files in the root tests/ directory for features, shared tests, and Supabase integration tests. Tests verify behavior via database operations and RLS policies — no UI testing. Triggers on: generate test, generate tests, create tests, write tests, test generation.

Display an overview of the Potenlab plugin ecosystem. Use for "info", "ecosystem", "about potenlab".

Orchestrates sequential project planning. First validates user intent via AskUserQuestion (MANDATORY), then spawns ui-ux-specialist to create ui-ux-plan.md, then tech-lead-specialist to create dev-plan.md from ui-ux-plan.md, then spawns frontend-specialist and backend-specialist in parallel to create their detailed plans from dev-plan.md, then spawns qa-specialist to create test-plan.md from dev-plan.md and backend-plan.md. Triggers on: plan, create plan, plan project, project plan, full plan, prd to plan.

Reviews and edits all existing plan files based on user feedback. Accepts feedback from arguments (/review-plan change X to Y) or via AskUserQuestion. Same sequential flow as /plan (ui-ux -> tech-lead -> frontend+backend parallel) but EDITS existing plans instead of creating from scratch. Also updates progress.json if it exists. Triggers on: review plan, edit plan, update plan, change plan, revise plan, adjust plan.

Runs ALL Vitest tests across /features, /tests, and /supabase directories. Executes tests via npx vitest run, collects results, and generates test.result.json (replaced on every run). Uses qa-specialist to analyze failures and provide actionable feedback. Triggers on: run test all, run all tests, test all, vitest all, run tests.

Runs Vitest tests for a specific phase chosen by the user. Reads test-plan.md to identify phases, asks user which phase to run via AskUserQuestion, executes only those test files, and generates test.result.json (replaced on every run). Uses qa-specialist to analyze failures. Triggers on: run test phase, test phase, run phase test, test specific phase.

Verifies and updates tests after code changes. Detects what changed (git diff, changes.json), compares against test-plan.md, then spawns qa-specialist agents to update test-plan.md and edit/create test files in /features, /tests, /supabase. Use after /developer or manual changes to keep tests in sync. Triggers on: verify test, verify tests, sync tests, update tests, check tests, test verify.