Agent· Database· Free· Claude, Coding agents· by Melvin Jones Repol
Claude Code subagent that rebuilds a missing Supabase schema from your app code. Writes one re-runnable SQL file with tables, FKs, RPCs, triggers, RLS, realtime, storage and seeds, marks every guess, and never touches a live database.
Supabase Schema Reconstructor
Claude Code subagent that rebuilds a missing Supabase schema from your app code. Writes one re-runnable SQL file with tables, FKs, RPCs, triggers, RLS, realtime, storage and seeds, marks every guess, and never touches a live database.
Design Consistency Refactor
Detect a project's existing design system from its code (tokens, colors, fonts, spacing, radius, components) and enforce it across the entire codebase by finding every page and component that doesn't follow it and refactoring them.
Engineering Principles
Write, refactor, and review code using core software engineering principles: KISS, DRY (applied with judgment), YAGNI ("just because you can doesn't mean you should"), minimal comments, single responsibility, clear naming, and consistency with the existing codebase.
WCAG Accessibility Auditor
Audit and fix UI for WCAG 2.2 AA accessibility. Use when the user asks if something is accessible, mentions a11y, WCAG, ADA, screen readers, keyboard navigation, color contrast, ARIA, or needs to pass an accessibility review. Works on code, screenshots, or design descriptions.
Microsoft Fluent 2
Design UI using Microsoft's Fluent 2 design system for Windows apps, Microsoft 365-style web apps, and Fluent UI React/Web Components.
Material Design 3
Design and build UI following Google's Material Design 3 (Material You).
UI design critique Gem that reviews screenshots against usability heuristics and your design system
A Gemini Gem for product designers: upload a screenshot or mockup with the user goal, and it gives a structured critique covering task flow, hierarchy, usability heuristics, accessibility basics and design system consistency, ranked by severity.
Pasted CSV analysis that checks the data before answering
Answers a business question about a CSV you paste or upload. It profiles the data first, states how each number was computed, separates findings from guesses and lists caveats, so the answer can be trusted.
Open-ended survey response coding with a reusable codebook
Builds a codebook from free-text survey answers, codes every response, and reports theme counts with representative quotes. You can reuse the codebook on the next survey so results stay comparable.
Design Consistency Refactor
Detect a project's existing design system from its code (tokens, colors, fonts, spacing, radius, components) and enforce it across the entire codebase by finding every page and component that doesn't follow it and refactoring them.
Engineering Principles
Write, refactor, and review code using core software engineering principles: KISS, DRY (applied with judgment), YAGNI ("just because you can doesn't mean you should"), minimal comments, single responsibility, clear naming, and consistency with the existing codebase.
WCAG Accessibility Auditor
Audit and fix UI for WCAG 2.2 AA accessibility. Use when the user asks if something is accessible, mentions a11y, WCAG, ADA, screen readers, keyboard navigation, color contrast, ARIA, or needs to pass an accessibility review. Works on code, screenshots, or design descriptions.
Microsoft Fluent 2
Design UI using Microsoft's Fluent 2 design system for Windows apps, Microsoft 365-style web apps, and Fluent UI React/Web Components.
Material Design 3
Design and build UI following Google's Material Design 3 (Material You).
Flaky test triage: reproduce, classify the cause, propose a fix
A Claude Code skill that takes one intermittently failing test, reruns it to measure the failure rate, classifies the likely cause (timing, shared state, clock, randomness, network, ordering) and proposes a targeted fix.
One-at-a-time dependency upgrade skill with changelog review
A Claude Code skill that upgrades a single dependency (or a short list, one by one), reads its release notes for breaking changes, updates affected call sites, and verifies with your typecheck, tests and build.
/release-notes: user-facing release notes between two git refs
A Claude Code slash command that turns the changes between two tags or commits into release notes written for end users and admins: highlights, improvements, fixes, and any action they must take to upgrade.
Incident debugging runbook skill: evidence first, read-only
A Claude Code skill for active production incidents that builds a timeline, ties symptoms to recent changes in the repo, ranks hypotheses, and suggests read-only checks, while keeping a running incident log you can paste into Slack or a postmortem.
Accessibility check for UI components against WCAG 2.2 AA
A Claude Code skill that reviews React, Vue, Svelte or HTML components for common WCAG 2.2 AA failures (labels, keyboard access, focus handling, ARIA misuse, contrast, target size) and reports or fixes them with file and line references.
OpenAPI 3.1 spec generator from your existing route handlers
A Claude Code skill that reads your HTTP routes, validators and response code, then writes or updates an OpenAPI 3.1 YAML file that matches what the server actually does, and lists anything it could not determine.
Supabase Schema Reconstructor
Claude Code subagent that rebuilds a missing Supabase schema from your app code. Writes one re-runnable SQL file with tables, FKs, RPCs, triggers, RLS, realtime, storage and seeds, marks every guess, and never touches a live database.
Design Consistency Refactor
Detect a project's existing design system from its code (tokens, colors, fonts, spacing, radius, components) and enforce it across the entire codebase by finding every page and component that doesn't follow it and refactoring them.
Engineering Principles
Write, refactor, and review code using core software engineering principles: KISS, DRY (applied with judgment), YAGNI ("just because you can doesn't mean you should"), minimal comments, single responsibility, clear naming, and consistency with the existing codebase.
WCAG Accessibility Auditor
Audit and fix UI for WCAG 2.2 AA accessibility. Use when the user asks if something is accessible, mentions a11y, WCAG, ADA, screen readers, keyboard navigation, color contrast, ARIA, or needs to pass an accessibility review. Works on code, screenshots, or design descriptions.
Microsoft Fluent 2
Design UI using Microsoft's Fluent 2 design system for Windows apps, Microsoft 365-style web apps, and Fluent UI React/Web Components.
Material Design 3
Design and build UI following Google's Material Design 3 (Material You).
UI design critique Gem that reviews screenshots against usability heuristics and your design system
A Gemini Gem for product designers: upload a screenshot or mockup with the user goal, and it gives a structured critique covering task flow, hierarchy, usability heuristics, accessibility basics and design system consistency, ranked by severity.
Pasted CSV analysis that checks the data before answering
Answers a business question about a CSV you paste or upload. It profiles the data first, states how each number was computed, separates findings from guesses and lists caveats, so the answer can be trusted.
Open-ended survey response coding with a reusable codebook
Builds a codebook from free-text survey answers, codes every response, and reports theme counts with representative quotes. You can reuse the codebook on the next survey so results stay comparable.
Flaky test triage: reproduce, classify the cause, propose a fix
A Claude Code skill that takes one intermittently failing test, reruns it to measure the failure rate, classifies the likely cause (timing, shared state, clock, randomness, network, ordering) and proposes a targeted fix.
Add a chat to your site that replies from your own pages. Jump in from the Inbox whenever you want.
Pasted CSV analysis that checks the data before answering
Answers a business question about a CSV you paste or upload. It profiles the data first, states how each number was computed, separates findings from guesses and lists caveats, so the answer can be trusted.
Open-ended survey response coding with a reusable codebook
Builds a codebook from free-text survey answers, codes every response, and reports theme counts with representative quotes. You can reuse the codebook on the next survey so results stay comparable.
Customer support reply drafter grounded in your policy text
Drafts a support reply from a ticket, account context and the policy excerpts you paste in. It answers only what your policies allow, flags anything it cannot promise, and adds internal notes for the agent.
Competitive analysis from your own source material, with gaps marked
Structures competitor research you have collected (websites, pricing pages, reviews, sales notes) into a comparison by job-to-be-done, a positioning map and a list of where you win, lose and do not know yet.
Post-call follow-up email from rough sales call notes
Turns messy notes from a sales or client call into a follow-up email that recaps what you heard, lists commitments on both sides and proposes one next step, plus a private list of your open items.
Socratic tutor that guides learners to answers without giving them away
Sets up a chat tutor that teaches any topic by asking one question at a time, diagnosing misconceptions from the learner's answers and giving hints in steps, so the learner does the thinking. Adapts to age and level.
Standard operating procedure writer from a rough process description
Turns a rambling explanation, voice-note transcript or screen-share notes of how a task is done into a clear SOP with purpose, roles, prerequisites, numbered steps, decision points, checks and a list of questions for the process owner.
Structured interview kit with behavioural questions and scoring guides
Builds an interview kit for one role and one interview stage: competencies tied to real outcomes, behavioural questions with follow-ups, a 1/3/5 scoring guide per competency and a timing plan, so every candidate gets a fair, comparable interview.
Customer-facing release notes from an internal change list
Converts an internal list of shipped changes (ticket titles, short descriptions) into release notes customers can read: grouped by what they can now do, written in plain language, with internal-only items left out.
Metric definition spec that removes ambiguity before you build a dashboard
Turns a loosely named metric ("active users", "churn", "time to resolution") into a precise definition: formula, inclusions, exclusions, time window, edge cases, owner and known pitfalls, ready for a data team to implement.
Three-touch cold email sequence personalised from one trigger
Writes a short, plain three-email outbound sequence built around one real trigger you found about the prospect, with subject lines and send timing. No fake familiarity, no invented proof points.
Supabase Schema Reconstructor
Claude Code subagent that rebuilds a missing Supabase schema from your app code. Writes one re-runnable SQL file with tables, FKs, RPCs, triggers, RLS, realtime, storage and seeds, marks every guess, and never touches a live database.
UI design critique Gem that reviews screenshots against usability heuristics and your design system
A Gemini Gem for product designers: upload a screenshot or mockup with the user goal, and it gives a structured critique covering task flow, hierarchy, usability heuristics, accessibility basics and design system consistency, ranked by severity.
RFP and security questionnaire response drafter that answers only from your approved answer library
A Claude Project for sales and bid teams: paste RFP questions and it drafts answers from your uploaded answer library and product docs, cites the source for each, and marks anything unsupported for a subject-matter expert.
CSV data quality profiler GPT that inspects an uploaded file with Python and writes a cleaning plan
A custom GPT for analysts and ops teams: upload a CSV or Excel export and it profiles every column with code (types, missing values, duplicates, outliers, inconsistent categories), reports issues with row counts, and proposes a cleaning plan you approve before any change.
Spreadsheet formula helper for Excel and Google Sheets that asks for your layout before writing a formula
A portable assistant for anyone working in Excel or Google Sheets: describe what you need and your column layout, and it writes the formula for your app and version, explains each part, lists edge cases, and gives a test row to check it.
Bias-aware hiring scorecard assistant for structured interviews and evidence-based ratings
A custom GPT for hiring managers and interviewers: builds a role scorecard with job-related competencies and behavioral anchors, writes structured interview questions, and turns interview notes into evidence-linked ratings while flagging non-job-related or biased language.
Project status report writer that turns raw notes into a RAG-rated weekly update
A portable assistant for project and program managers: paste messy notes, tickets or standup logs and get a consistent weekly status report with an honest red/amber/green rating, risks with owners, decisions needed and next steps.
New-hire onboarding buddy that answers from your handbook and says when to ask a person
A Claude Project for people teams and managers: upload your handbook and onboarding docs, and new hires get answers with the source section named, a first-weeks checklist, and a clear handoff to HR or their manager when the docs are silent.
Security reviewer subagent for branch diffs
A Claude Code subagent that reviews the changes on a branch for exploitable security issues (injection, broken access control, secrets, SSRF, path traversal, unsafe deserialization, weak crypto) and reports each with an attack scenario and fix.
Claim checker GPT that extracts factual claims from a draft and verifies each against sources
A custom GPT for writers, editors and comms teams: paste a draft and it lists every checkable claim, searches for supporting or contradicting sources with web search, and returns a verdict table with links, so you fix problems before publishing.
TypeScript strict-mode migration subagent
A Claude Code subagent that turns on TypeScript strict checks one flag at a time and fixes the resulting errors directory by directory with real types, without falling back to `any` or blanket `@ts-ignore`.
Prompt regression runner for the Anthropic API with a PR workflow
A Python script that runs JSONL test cases through the Anthropic Messages API and checks each reply with contains, regex, length and JSON rules, plus a GitHub Actions workflow that runs it on PRs touching prompts.
Jupyter notebook output stripper and CI check
A standard-library Python script that clears outputs and execution counts from .ipynb files, plus a GitHub Actions check that fails pull requests containing notebooks with outputs, so API responses and sample data do not get committed.
Auto-label pull requests that change prompts, evals or agent config
An actions/labeler v5 configuration and workflow that labels PRs touching prompts, eval sets, training data, or AI assistant config files such as CLAUDE.md, AGENTS.md, .mcp.json and .cursorrules.
CI secret scan for AI provider API keys in code, notebooks and fixtures
A dependency-free Python scanner and GitHub Actions workflow that blocks pull requests containing Anthropic, OpenAI, AWS, GitHub, Google, Slack or Hugging Face keys, or private keys, with masked output and an allowlist file.
GitHub Actions check that JSON, JSONL and YAML files parse
A CI workflow and a small Python script that fail a pull request when any JSON, JSONL or YAML file in the repo is malformed or a JSON object repeats a key. Built for repos full of prompt configs, eval sets and agent settings.
Claude Code hooks that block destructive shell commands and protect secret files
A drop-in .claude/settings.json with two PreToolUse hooks: one denies dangerous Bash commands (recursive deletes of root or home, force pushes to main, disk wipes, piping curl to a shell), the other blocks edits to .env files, keys and lockfiles.
Prompt file linter with front matter, variable and tag checks
A Python linter and CI workflow for prompt templates in prompts/. It catches undeclared {{variables}}, unclosed XML-style tags, missing front matter, duplicate prompt names and invisible characters before they reach a model.
LLM-as-judge grader for model outputs using the OpenAI API (Node.js)
A Node.js script that grades a JSONL file of model outputs against your written rubric with an OpenAI judge model, returning a 1-5 score and reasons per case, a pass rate, and CI-friendly exit codes.
Markdown relative link and anchor checker for docs and prompt repos
A dependency-free Python script and GitHub Actions workflow that find broken relative links and missing #heading anchors in Markdown files, without making any network calls.
Claude Code hook that formats and lints each file after Claude edits it
A PostToolUse hook for .claude/settings.json that runs the right formatter on every file Claude writes or edits (Prettier, Ruff, gofmt, rustfmt, Pint) and sends remaining lint errors back to Claude to fix.
Prompt injection regression test cases for LLM assistants
Ten ready-to-run JSONL test cases covering direct, role-play, encoding, indirect (email, web page), fake-admin and markdown-image exfiltration injections, using a canary string so leaks are detected by simple checks.
AGENTS.md template for Next.js App Router projects (TypeScript)
An AGENTS.md for Next.js App Router codebases that tells any coding agent your commands, server and client component rules, data-fetching and caching conventions, server action safety, and what not to touch.
Prompt template renderer that fails on missing variables (Node.js)
A zero-dependency Node.js script that fills {{variable}} placeholders in a prompt file from JSON, supports nested keys and lists, and exits non-zero on missing or unused variables so broken prompts never reach a model.
Postmortem template for AI and LLM feature incidents
A blameless incident postmortem template adapted for AI features: records model and prompt versions, user and data impact, agent actions taken, an AI-specific failure classification, layered root cause questions, and eval-focused action items.
GitHub Copilot instructions for FastAPI, SQLAlchemy 2.0 and Pydantic v2
A .github/copilot-instructions.md that steers Copilot toward current FastAPI patterns: SQLAlchemy 2.0 select() style, Pydantic v2 APIs, thin routers with service layers, per-record authorization checks, and pytest with dependency overrides.
CLAUDE.md template for Laravel applications
A ready-to-fill CLAUDE.md for Laravel projects that gives Claude Code the commands, conventions and guardrails it needs: tests with Pest or PHPUnit, Pint, migrations, Form Requests, policies, queues and config caching rules.
Token and cost estimator script for prompts and batch jobs
A Python script that approximates token counts for prompt files and projects the cost of a batch of requests from prices you supply. Clearly labeled as an estimate, with optional tiktoken counting.
JSON Schema for support ticket triage structured output
A strict JSON Schema (draft 2020-12) for classifying an inbound support ticket: category, priority, sentiment, language, summary, extracted IDs, human-handoff flag and routing queue. Every field is required and described.
Cursor rules file for Next.js App Router, TypeScript and Tailwind
A .cursorrules file that keeps AI-generated code in a Next.js App Router project consistent: server-first components, parallel data fetching, validated Server Actions, Zod at trust boundaries, Tailwind-only styling and safe env handling.
Regex PII redactor for prompts, logs and datasets
A dependency-free Python script that replaces emails, phone numbers, US SSNs, Luhn-valid card numbers, IP addresses, IBANs and bearer tokens with placeholders before text goes to a model or into a dataset. Its limits are stated up front.
Project .mcp.json for Claude Code with filesystem, memory, fetch and sequential-thinking servers
A project-scoped .mcp.json that connects Claude Code to four official reference MCP servers: filesystem access to chosen folders, a persistent knowledge-graph memory, web page fetching, and a structured step-by-step thinking tool.
CSV to chat-format JSONL converter for fine-tuning datasets
A Python script that turns a spreadsheet export of prompt and ideal-answer pairs into chat-format JSONL (one messages array per line) for fine-tuning, with column mapping, a shared or per-row system message, dedupe, and atomic output.
Supabase Schema Reconstructor
Claude Code subagent that rebuilds a missing Supabase schema from your app code. Writes one re-runnable SQL file with tables, FKs, RPCs, triggers, RLS, realtime, storage and seeds, marks every guess, and never touches a live database.
Microsoft Fluent 2
Design UI using Microsoft's Fluent 2 design system for Windows apps, Microsoft 365-style web apps, and Fluent UI React/Web Components.
Material Design 3
Design and build UI following Google's Material Design 3 (Material You).
Design Consistency Refactor
Detect a project's existing design system from its code (tokens, colors, fonts, spacing, radius, components) and enforce it across the entire codebase by finding every page and component that doesn't follow it and refactoring them.
Engineering Principles
Write, refactor, and review code using core software engineering principles: KISS, DRY (applied with judgment), YAGNI ("just because you can doesn't mean you should"), minimal comments, single responsibility, clear naming, and consistency with the existing codebase.
WCAG Accessibility Auditor
Audit and fix UI for WCAG 2.2 AA accessibility. Use when the user asks if something is accessible, mentions a11y, WCAG, ADA, screen readers, keyboard navigation, color contrast, ARIA, or needs to pass an accessibility review. Works on code, screenshots, or design descriptions.
UI design critique Gem that reviews screenshots against usability heuristics and your design system
A Gemini Gem for product designers: upload a screenshot or mockup with the user goal, and it gives a structured critique covering task flow, hierarchy, usability heuristics, accessibility basics and design system consistency, ranked by severity.
Pasted CSV analysis that checks the data before answering
Answers a business question about a CSV you paste or upload. It profiles the data first, states how each number was computed, separates findings from guesses and lists caveats, so the answer can be trusted.
Open-ended survey response coding with a reusable codebook
Builds a codebook from free-text survey answers, codes every response, and reports theme counts with representative quotes. You can reuse the codebook on the next survey so results stay comparable.
A creator marketplace for AI skills, prompts, agents, automations, and files. Plus an AI live chat for your website, and an MCP server for your AI tools.
New features as they ship, AI live chat improvements, and the best new listings in the catalog.