assistants

Copy processed MP3s and WAVs to NAS storage. Use after Apple Music import or when re-archiving corrected files.

Move processed ZIPs to archive and clean up extraction folders. Use after album processing is complete.

Copy finished MP3s into Apple Music via the auto-import folder and verify the album lands correctly. Use once metadata is finalized and the user wants tracks added to their library — e.g. "add this to Apple Music" or "import these tracks" — even if they don't say "auto-import" explicitly. Also use to diagnose a bad import, e.g. "it showed up as separate tracks instead of one album."

Inspect and update MP3 tags: genre, artist, album, track count, compilation flag. Use when checking or fixing music file metadata — including diagnostic questions like "why do these show up as separate tracks" or "is this folder's metadata consistent," not just explicit tag-editing requests.

End-to-end processing of downloaded music purchases: extract ZIPs, verify metadata, import to Apple Music, archive to NAS, clean up. Use when processing new music downloads.

Reclaim local disk by offloading cloud-backed Apple Music downloads. Audit iCloud status, build an offload playlist that protects your DJ crates and lossless files, then Remove Download. Use when the Mac is low on disk.

Find music release ZIPs in Downloads and classify each as MP3 or WAV by inspecting archive contents (never filenames), matching pairs by release name. Use to see what's available to process — e.g. "what did I just buy" or "what's in my downloads" — even if the user doesn't mention ZIPs or file formats directly.

Check that all media-management paths (Downloads, Apple Music, NAS archive, rekordbox-mcp) are resolvable via env var or config.json, and help the user fix any that are missing. Use when paths seem not configured, when a skill fails to resolve a path, when explicitly asked to set up/configure the plugin, or when asked where its config lives or how path resolution works.

Split audio tracks exceeding 78 minutes at silence points with crossfades. Use when tracks are too long for Apple Music.

Swift 6 language and tooling — concurrency model, Sendable, actors, XCTest basics, and xcodebuild. For SwiftUI architecture and view-layer testing, see `swiftui-lang`.

SwiftUI architecture and testing — ViewModel-first extraction, protocol-based DI, coverage targets by layer, ViewInspector, snapshot testing, how to safely pass closures to Apple SDK APIs inside @MainActor types, and common anti-patterns.

A no-op skill used as the include target for ynh E2E include/delegate tests.

Compose a structured feedback report for a harness artifact problem — gathers required context and formats it clearly for submission as a GitHub issue, PR description, or message.

Fetch current documentation from LLM vendor sources — given a vendor or artifact type, retrieves up-to-date content from canonical URLs for comparison against stored references.

Maintain the vendor gap table — add new gaps, update existing entries, and mark gaps as resolved when vendor support improves.

Validate, diagnose, and understand hook configurations across Claude Code, Cursor, and Codex — event coverage, hook types, format differences, and why hooks may not fire.

Determine where a harness artifact originated — tries multiple provenance strategies in order and returns the source repo, path, and ref when found.

Validate, diagnose, and understand MCP server declarations and runtime behavior across Claude Code, Cursor, and Codex — transport types, declaration format, and vendor support.

Audit an entire plugin directory for cross-vendor readiness — walks all artifacts (skills, agents, MCP, hooks, startup context) and reports issues by vendor and priority.

Validate, diagnose, and understand Agent Skills (SKILL.md) across Claude Code, Cursor, and Codex — spec compliance, vendor-specific loading behavior, and known quirks.

Validate and diagnose startup context files — CLAUDE.md, AGENTS.md, rules, and .cursorrules — that inject instructions into the model context at session start across Claude Code, Cursor, and Codex.

Validate, diagnose, and understand subagent definitions (agents/*.md) — frontmatter format, delegation model, and vendor support across Claude Code, Cursor, and Codex.

Select the right feedback channel and submit — GitHub issue, GitHub PR, email, or message — based on provenance, contributor status, and local checkout availability.

Master cross-vendor reference for LLM vendor harness formats — complete format mapping tables, documentation URLs, known gaps, and the workflow for keeping references current.

Tag-driven release automation — versioning, release types, and monitoring. Never create releases manually.

Set up and harden a public GitHub repository — repo settings, security, branch protection, templates, CI, and dependabot.

Four-gate quality model — build, lint, format, test — that must all pass before any commit.

Gitflow branching model — develop/main strategy, branch naming, PR rules, and post-merge cleanup.

Push committed work, open a PR targeting the correct base branch, monitor CI, and merge when ready.

Run the full code quality pipeline — lint, format, test, and build — with zero tolerance for errors.

GitHub Actions workflows that implement the Gitflow model — CI gating, branch protection enforcement, and tag-triggered release.

Conventional commit message format, PR description template, and PR title conventions.

GitHub operations via gh CLI — use this instead of GitHub MCP server tools for all PR, issue, release, and CI work.

Author agent skills conforming to the agentskills.io spec — SKILL.md format, frontmatter fields, directory layout, reference docs, and progressive disclosure patterns.

Backend development approach — API design, data flow, error handling, and service architecture patterns.

Interactive debugging session discipline — visibility-first debugging, safe git workflow, and pair programming patterns.

Project setup, directory scaffolding, and initial configuration for new software projects.

Structured code review covering correctness, security, performance, and maintainability.

Security-first development — input validation, authentication, secrets management, and OWASP top 10 prevention.

UI development approach — component architecture, state management, accessibility, and user experience patterns.