martonpaulo
UserAgent Skills for coding agents: software engineering workflows and architecture tools. Personalized forks, with every upstream author credited.
Categories
Indexed Skills (27)
apple-docs
Research authoritative, version-aware Apple platform documentation for Swift, SwiftUI, UIKit, AppKit, Foundation, macOS, iOS, Xcode, Apple SDKs, Swift Evolution, compiler behavior, HIG, WWDC, signing, entitlements, capabilities, sandboxing, notarization, privacy manifests, distribution, and platform release changes. Use when implementation correctness depends on Apple API behavior, lifecycle, availability, deprecation, SDK or Xcode version, deployment target, platform conventions, or official guidance. In existing projects, detect Xcode, Swift, SDK, platform, dependencies, and deployment targets first. Prefer official Apple and Swift sources and clearly label forum discussions and community notes as secondary. Do not use for purely algorithmic Swift questions or non-Apple documentation.
bug-hunter
Audit an existing codebase for verified functional, logic, runtime, concurrency, data-integrity, contract, and security bugs without fixing them. Use only when the user explicitly invokes $bug-hunter for adversarial bug finding in a repository, path, staged change, or branch diff. Not for ordinary code review, style issues, missing tests, speculative hardening, implementation, auto-fix, or remediation work.
code-review
Perform the Validate SDD phase for exactly one pull request. Triggers when the user provides a PR number/URL to review.
context7
Fetch current documentation snippets and code examples for an indexed third-party library, framework, SDK, or CLI tool from the Context7 index, using the ctx7 command line tool. Use when implementation depends on a library's present API surface, configuration options, or usage patterns, training data may be stale, and a fast indexed lookup is enough. Defer Apple platform documentation to apple-docs. Escalate to deep-docs when the conclusion must be traced to an official versioned source, when the library is not indexed, or when compatibility, deprecation, or migration has to be proven. Do not use for the project's own code, general web research, product comparisons, build-versus-buy decisions, or managing skills from the Context7 registry.
deep-docs
Perform authoritative, version-aware documentation research for software, frameworks, SDKs, libraries, APIs, CLIs, databases, developer tools, and platforms used in the current task or project. Detect the exact product and installed version from project files or safe local commands, then search official versioned documentation, release notes, API specifications, official source repositories, local CLI manuals, and explicitly configured read-only documentation sources. Use when implementation correctness depends on precise behavior, version compatibility, availability, deprecation, configuration, migration, or documented constraints. Prefer primary sources and return source-linked evidence. Defer Apple development documentation to apple-docs. Do not use for general web research, product recommendations, build-versus-buy decisions, debugging without a documentation question, or broad architecture decisions.
disk-cleaner
Audit a personal computer for reclaimable disk space and, only after explicit approval, clean it. Covers caches, logs, build artifacts, dependency stores, duplicates, stale downloads, package-manager leftovers, simulator and container data, and residue from uninstalled apps. Always starts read-only, classifies every candidate by risk and recreatability, and never deletes anything without a confirmed decision. Use when the user asks what is filling their disk, wants a cleanup plan, or asks to free space on their machine. Do not use for cleaning a project's working tree during normal development, for diagnosing application bugs, or for any repository-level task.
product-audit
Audit an existing product interface for evidence-backed UI, UX, accessibility, and copy findings at low, medium, or high depth. Use only when the user explicitly invokes $product-audit for a read-only review of a screen, flow, feature, live product, design artifact, or implementation. Not for implementing fixes, redesigning the product, writing back to design tools, or changing code.
handoff
Write a compact durable handoff containing the exact context, decisions, modified files, verification, risks, and next action another agent or later session needs to continue safely.
architecture-review
Review an existing codebase for high-value architectural improvements, using concrete code evidence to identify weak boundaries, shallow modules, duplication, coupling, scattered responsibilities, and unstable seams.
backlog-curator
Maintain the quality, consistency, and implementability of the repository's GitHub Issue backlog. Detects duplicates, obsolete issues, and delegates missing SDD phases to capture-issue or plan-issue. Closes with a Mermaid graph of the blocking order across open issues.
capture-issue
Convert a newly reported or discovered bug, defect, problem, or feature request into one canonical, implementation-ready GitHub Issue containing the Specify and Clarify phases. Do not trigger when the defect will be fixed immediately, an adequate issue exists, the user only wants an explanation, or the request is already in later phases.
dont-reinvent-the-wheel
Evaluate whether a specific requested software capability should use an existing project feature, native platform capability, maintained dependency, open-source project, external service, hybrid approach, or custom implementation. Use for explicit build-versus-reuse decisions or when evaluating replacements for one identified custom capability. Do not use for general research, architecture reviews, codebase-wide audits, implementation work, debugging, or requirements interviews.
grey-market
Find digital products (accounts, licenses, subscriptions, gift cards, templates, courses) sold far below the official price in regional markets such as China, India, Brazil, Russia, Southeast Asia and Eastern Europe, by searching discussion communities, forums, messaging groups, and regional marketplaces instead of ordinary search engines. Use when the user asks for cheap keys or licenses, shared accounts, regional offers, loot deals, Brazilian PIX-friendly sellers, Turkish or Argentine gift cards, subscription splitting, or names a grey-market source such as Plati, 52pojie, DesiDime, GGMAX, or Desapego Games. Do not use for ordinary shopping research, for pirated software, or for anything that requires making a purchase.
implement-issue
Perform the Implement SDD phase for exactly one prepared issue. Triggers when the user provides an issue number/URL to implement.
module-design
Design or improve module boundaries, interfaces, dependency direction, cohesion, coupling, test seams, and responsibility ownership when a code change exposes unclear or unstable structure.
plan-issue
Complete the Plan and Tasks SDD phases for exactly one GitHub Issue. Triggers when the user provides an issue number/URL to plan, or when delegated by backlog-curator.
prototype
Build a small disposable experiment when executing code is the fastest reliable way to answer a concrete technical, behavioral, integration, state-management, or UI question. Do not use for normal production implementation.
setup-agent-docs
Configure lightweight per-repository conventions for optional domain glossaries, ADRs, research notes, handoffs, and disposable prototypes. Use when the user directly requests these artifact paths or when setup-project delegates this step after establishing root guidance. Do not use for complete project setup, ordinary documentation edits, personal-skill configuration, trackers, tickets, labels, or backlog workflows.
setup-project
Initialize or align one of the owner's repositories with its durable project identity, Git and release policies, root AGENTS.md rules, CLAUDE.md alias, public documentation, licensing, ignore rules, GitHub metadata, and applicable CI or release foundations. Use only when the user explicitly asks to set up a new personal project or adopt this complete baseline in an existing repository; do not use for ordinary implementation, a single documentation edit, or artifact-path configuration alone.
skill-authoring
Create, review, or simplify Agent Skills with precise triggers, focused workflows, explicit safety boundaries, useful completion criteria, valid frontmatter, and minimal unnecessary context.
debug
Diagnose a non-trivial bug through reproduction, evidence gathering, hypothesis testing, root-cause analysis, a minimal fix, and regression verification. Use when the cause is not already obvious.
domain-model
Clarify ambiguous domain terminology, states, rules, entities, and relationships when code, documentation, or requested behavior use inconsistent or overloaded language.
grill-and-document
Start a focused one-question-at-a-time interview that also preserves resolved domain vocabulary and genuinely consequential architectural decisions without creating tickets or implementing the plan.
grilling
Interview the user one decision at a time to resolve material uncertainty in a plan or design before implementation. Use only when unresolved choices could significantly affect behavior, scope, architecture, data, security, compatibility, or user experience.
research
Research a technical or product question using high-trust primary sources when a decision depends on current documentation, external evidence, compatibility details, standards, or verified behavior.
resolve-conflicts
Resolve an in-progress Git merge, rebase, or cherry-pick conflict by reconstructing the intent of both sides, preserving compatible behavior, and validating the resulting code and history.
grill
Start a focused one-question-at-a-time interview to pressure-test a plan or design without creating documentation or implementing it.
Bio shown is the top-scored skill's repo description as a fallback — real GitHub bios land in a future update.