Skills catalog

Skills are SKILL.md files synced into .claude/skills/, and also into .codex/skills/ and .agents/skills/ when [project].tools names codex or antigravity. SKILL.md is a standard those tools share — only the directory differs — so the same package is copied verbatim into each. core skills apply everywhere; profile skills only reach projects with that profile enabled. Source lives at core/skills/<name>/SKILL.md or profiles/<profile>/skills/<name>/SKILL.md in the lacquer repo.

Every Source links somewhere: a skill’s own file in this repo (authored here, or with no further attributable external source), or the named upstream it was vendored from (copied in, then adapted to this repo’s conventions) — see Notes on vendored skills below each table for anything a short link can’t capture.

core

claudemd-review, handoff, and private-repo-search are this repo’s own implementations of three commands from the dx plugin (ykdojo/claude-code-tips) — written from scratch, not copied, so they follow this repo’s own conventions and install without the plugin’s other five commands, which don’t fit this fleet (/dx:gha overlaps github-ci-fix; /dx:half-clone needs to rewrite the live session transcript, not something a SKILL.md can do; /dx:reddit-fetch//dx:hn-summarize//dx:version-check are generic productivity, not fleet engineering). See docs/references.md.

Skill	Fires when	Source
advisor-checkpoint	Consulting a stronger model mid-task for a second opinion — stuck, before committing to an approach, or before declaring a non-trivial task done.	advisor-checkpoint
antigravity-rescue	Stuck, and wanting a second opinion from Google’s Antigravity CLI (agy) via a real subprocess — forwards the task and returns its output verbatim.	antigravity-rescue
caveman	User says “caveman mode”, “less tokens”, “be brief”, or invokes /caveman — token-efficient terse replies.	caveman
claude-rescue	Running under Codex CLI, stuck, and wanting a second opinion from Claude Code via a real subprocess — forwards the task and returns its output verbatim.	openai/codex-plugin-cc (codex-rescue, reversed)
claudemd-review	Corrected on something a CLAUDE.md file should already have covered, explicitly asked to review CLAUDE.md, or after a session with a lot of back-and-forth — reviews this conversation and proposes a bounded edit.	claudemd-review — own version of the dx plugin’s /dx:review-claudemd (see notes below)
evaluator-optimizer	Generate a solution, evaluate against explicit pass/fail criteria, refine, repeat until it passes or a round cap is hit.	evaluator-optimizer
exploring-an-idea	The very start of a new product idea — before a PRD, PCD, repo, or name exists, or when someone asks for a PRD/PCD from scratch.	exploring-an-idea
github-ci-fix	PR checks fail, CI is red, or a GitHub Actions workflow breaks — inspects failing checks via gh, pulls logs, triages flaky-vs-real and scopes the breaking commit, scopes external checks, builds a fix plan.	github-ci-fix — flakiness/breaking-commit triage folded in to cover the dx plugin’s /dx:gha without a second competing skill
github-issue-fix-flow	Given a GitHub issue number: implement a fix, run builds/tests, commit with a closing message, push.	github-issue-fix-flow
handoff	Before a session ends, before a context-heavy interruption (SSH/tmux disconnect, low context), or a worktree is handed to another worker — writes progress, what worked/failed, and the next concrete step.	handoff — own version of the dx plugin’s /dx:handoff (see notes below)
manager-loop	Running a large batch of independent work items (fleet-wide harvest, multi-repo sync-down, overnight backlog) as a persistent coordinator.	manager-loop
nameplate-attention	Grabbing the human’s attention (topmost message card + pulsating screen borders) before a password-manager prompt or whenever blocked on them.	steipete/nameplate
private-repo-search	Finding code or a past implementation across GitHub repos — including private ones — via gh search code, reaching further than grep/Glob can.	private-repo-search — own version of the dx plugin’s /dx:private-github-search (see notes below)
security-review	Adversarial security review of a branch diff, PR, or working tree before merge.	security-review
skill-authoring-standard	Writing or reviewing a SKILL.md — the bar it must clear: tight trigger-oriented frontmatter, single responsibility, no padding.	skill-authoring-standard
skill-tuning-loop	Empirically testing whether a skill needs an edit — mines session transcripts for friction (or audits against skill-authoring-standard when evidence is thin), proposes a bounded fix, validates against held-out cases before it ships.	skill-tuning-loop
working-with-lacquer	Working in a lacquer-managed project and the management itself comes up — the drift audit in Detect changed paths fails, lacquer audit exits 3/4/6, a synced config reverts on the next sync, or a baseline can’t be met yet.	working-with-lacquer

marketing

Vendored wholesale from coreyhaines31/marketingskills (MIT; the upstream licence is kept at profiles/marketing/skills/LICENSE-upstream-marketingskills). They used to sit in core, on the reasoning that they apply to any project with a public-facing app or site, not just an iOS one — but core ships to every project regardless of what it does, and most projects in this fleet are apps with no marketing surface of their own. They now ship only to a component that deliberately declares the marketing profile (see Profiles that ship in the README) — never auto-detected, since there is no marketing “stack” on disk to find. Two were edited on the way in rather than copied blind: video used npm install, which the web profile bans in favour of pnpm against a committed lockfile, and cold-email used {{FirstName}}-shaped mail-merge placeholders that collide with lacquer’s own {{TOKEN}} substitution syntax and made sync correctly refuse to render them.

Skill	Fires when
ab-testing	Planning, designing, or implementing an A/B test, split test, or a growth experimentation program.
ad-creative	Generating or scaling ad creative — headlines, descriptions, primary text, full variations — for any paid platform.
ads	Paid advertising campaigns on Google, Meta, LinkedIn, X, or another ad platform; PPC, ROAS, CPA, retargeting.
ai-seo	Optimizing content to be cited by LLMs and appear in AI-generated answers (AEO / GEO / LLMO).
analytics	Setting up, improving, or auditing analytics and measurement — GA4, event tracking, UTMs, GTM.
aso	Auditing or optimizing an App Store or Google Play listing.
attribution	Working out which marketing actually drives revenue, choosing an attribution model, or reconciling numbers that disagree across tools.
churn-prevention	Reducing churn — cancellation flows, save offers, dunning, failed-payment recovery, retention.
co-marketing	Finding co-marketing partners and planning joint or cross-promotional campaigns.
cold-email	Writing B2B cold outreach emails and follow-up sequences.
community-marketing	Building or growing a Discord/Slack/forum community and turning it into word-of-mouth.
competitor-profiling	Researching and profiling competitors from their URLs.
competitors	Building competitor comparison, “vs”, and alternative pages.
content-strategy	Deciding what content to create — topic clusters, blog strategy, content planning.
copy-editing	Editing, reviewing, or refreshing existing marketing copy.
copywriting	Writing or rewriting marketing copy for a page — homepage, landing, pricing, feature, about.
cro	Increasing conversions on a marketing page or a form.
customer-research	Running or synthesizing customer research — interviews, surveys, support tickets, ICP.
directory-submissions	Submitting a product to startup / SaaS / AI directories for backlinks and discovery.
emails	Building an email sequence, drip campaign, or lifecycle flow.
events	Planning, running, sponsoring, or speaking at events, and getting pipeline out of them.
free-tools	Planning or building a free tool as marketing (engineering as marketing).
image	Creating or optimizing marketing images — blog heroes, social graphics, mockups, brand assets.
influencer-marketing	Running influencer, creator, or ambassador partnerships, including disclosure compliance.
launch	Planning a product launch, feature announcement, or release strategy — Product Hunt, beta, waitlist.
lead-magnets	Creating or optimizing a lead magnet for email capture.
marketing-council	Wanting several expert perspectives on one marketing question — a simulated board of legendary marketers.
marketing-ideas	Needing marketing, growth, or promotion ideas for a software product.
marketing-loops	Setting up a recurring, self-running marketing workflow an agent runs on a cadence.
marketing-plan	Needing a full marketing or go-to-market plan — AARRR, 90-day, 12-month.
marketing-psychology	Applying behavioral science, mental models, or persuasion principles to marketing.
offers	Designing or improving the offer itself — value framing, bonuses, guarantees, naming, payment structure.
onboarding	Optimizing post-signup onboarding, activation, first-run experience, time-to-value.
paywalls	Creating or optimizing in-app paywalls, upgrade screens, upsells, feature gates.
popups	Creating or optimizing popups, modals, slide-ins, or banners for conversion.
pricing	Pricing decisions, packaging, and monetization strategy — tiers, freemium, trials, willingness to pay.
product-marketing	Creating or updating the product marketing context document — positioning, ICP, target audience.
programmatic-seo	Creating SEO pages at scale from templates and data.
prospecting	Finding, qualifying, and building a prospect list.
public-relations	Earned media, press coverage, journalist outreach, media strategy.
referrals	Creating, optimizing, or analyzing a referral, affiliate, or word-of-mouth program.
revops	Revenue operations — lead lifecycle, scoring, routing, marketing-to-sales handoff, pipeline stages.
sales-enablement	Sales collateral — pitch decks, one-pagers, objection handling, demo scripts.
schema	Adding or fixing schema markup and structured data — JSON-LD, rich snippets.
seo-audit	Auditing or diagnosing SEO issues — technical SEO, on-page, a traffic drop.
signup	Optimizing signup, registration, and trial activation flows.
site-architecture	Planning page hierarchy, navigation, URL structure, and internal linking.
sms	Planning or optimizing SMS/MMS marketing — welcome, abandoned cart, win-back, transactional.
social	Creating, scheduling, or optimizing social content, and social listening / engagement triage.
video	Producing video content with AI tools or programmatic frameworks (Remotion, Veo, Sora, Runway, …).

ios

Several skills in this profile are vendored from public, per-topic “Agent Skill” repos by AvdLee (Antoine van der Lee — SwiftLee); others are original to this repo. See the notes below the table.

Skill	Fires when	Source
app-store-screenshots	Capturing App Store/TestFlight screenshots from the simulator at native resolution and uploading via helm-asc.	app-store-screenshots
core-data-expert	Setting up a Core Data stack, debugging threading/merge conflicts, planning a migration, integrating CloudKit sync, or diagnosing performance/memory issues.	AvdLee/Core-Data-Agent-Skill
helm-asc	Driving App Store Connect through the Helm CLI — app discovery, versions, localizations, screenshots, TestFlight groups, builds, IAPs and subscriptions, review replies and submissions.	helm-asc — copied from the already-working local skill rather than reconstructed from vendor docs (see notes below)
ios-debugger-agent	Running an iOS app, interacting with the simulator UI, capturing logs, or diagnosing runtime behavior.	ios-debugger-agent
ios-performance-battery-patterns	Touching widgets, animations, networking configuration, or background/observer cleanup — battery and performance patterns.	ios-performance-battery-patterns — split out of CLAUDE.ios.md to keep it under the ~200-line guideline
ios-secrets-setup	Wiring a new app-runtime service key (RevenueCat, Aptabase, …) into Secrets.xcconfig, project.yml, and Info.plist.	ios-secrets-setup — split out of CLAUDE.ios.md to keep it under the ~200-line guideline
macos-ci-recipes	Adding macOS build/test/lint CI to a project on the ios profile (macOS-only or hybrid with an iOS target).	macos-ci-recipes; adapted from foxtrot’s and uniform’s own CI (own fleet)
macos-development	AppKit interop/modernization, App Sandbox & file-access entitlements, menu bar commands, multi-window scenes, the Settings scene, ScreenCaptureKit, macOS-specific SwiftUI differences, notarized Developer ID distribution.	adapted from CharlesWiltgen/Axiom’s axiom-macos skill (MIT); see docs/references.md
native-app-profiling	Profiling native macOS/iOS apps for CPU hotspots, hangs, and hitches via Instruments traces.	native-app-profiling
release-app-store-changelog	Generating user-facing App Store “What’s New” release notes from git history.	release-app-store-changelog
release-macos-spm-packaging	Scaffolding, building, and packaging SwiftPM-based macOS apps without an Xcode project.	release-macos-spm-packaging; harvested from kit (own fleet)
swift-concurrency	Diagnosing data races, converting callbacks to async/await, actor isolation, Sendable conformance, Swift 6 migration.	AvdLee/Swift-Concurrency-Agent-Skill
swift-testing-expert	Writing or modernizing Swift Testing suites — #expect/#require, traits/tags, parameterized tests, XCTest migration.	AvdLee/Swift-Testing-Agent-Skill
swift-testing-wait-until	A test needs to wait for an async state change instead of a fixed delay — the no_task_sleep_in_tests lint rule bans arbitrary Task.sleep.	swift-testing-wait-until — split out of CLAUDE.ios.md to keep it under the ~200-line guideline
swiftui-expert-skill	Writing, reviewing, or improving SwiftUI code — state management, view composition, performance, Liquid Glass adoption; also analyzes Instruments .trace files.	AvdLee/SwiftUI-Agent-Skill
swiftui-liquid-glass	Building or reviewing SwiftUI UI against the iOS 26+ Liquid Glass API, the default design from the iOS 27 SDK.	split from AvdLee/SwiftUI-Agent-Skill’s reference material
swiftui-performance-audit	Diagnosing slow rendering, janky scrolling, high CPU/memory, or excessive view updates in SwiftUI.	split from AvdLee/SwiftUI-Agent-Skill’s reference material
swiftui-ui-patterns	Creating or refactoring SwiftUI UI, tab architecture, screen composition.	swiftui-ui-patterns
swiftui-view-refactor	Cleaning up a SwiftUI view’s structure, dependency injection, and Observation usage.	swiftui-view-refactor
update-swiftui-apis	Scanning Apple’s SwiftUI docs for deprecated APIs and refreshing the swiftui-expert-skill with replacements (requires the Sosumi MCP).	update-swiftui-apis
url-validation-security	Validating a user-provided or externally-sourced URL before it reaches AVPlayer, URLSession, or a WKWebView.	url-validation-security — split out of CLAUDE.ios.md to keep it under the ~200-line guideline
watchos-development	watchOS app structure, Watch Connectivity, complications/Smart Stack, controls & Live Activities on watch, background refresh/networking limits, WatchKit/ClockKit modernization.	adapted from CharlesWiltgen/Axiom’s axiom-watchos skill (MIT); see docs/references.md
xcode-build-benchmark	Benchmarking clean/incremental Xcode builds with repeatable inputs and timestamped artifacts.	AvdLee/Xcode-Build-Optimization-Agent-Skill
xcode-build-fixer	Applying an approved Xcode build optimization plan, then re-benchmarking.	AvdLee/Xcode-Build-Optimization-Agent-Skill
xcode-build-orchestrator	End-to-end build optimization: benchmark, run specialist analyzers, prioritize, get approval, delegate fixes, re-benchmark.	AvdLee/Xcode-Build-Optimization-Agent-Skill
xcode-compilation-analyzer	Analyzing Swift compile hotspots and type-checking cost from build timing summaries.	AvdLee/Xcode-Build-Optimization-Agent-Skill
xcode-disk-cleanup	Low disk space on a Mac that builds — auditing DerivedData, simulator runtimes and devices, archives, dSYMs, DeviceSupport and stale caches against evidence of what still uses them, then cleaning behind an approval gate.	AvdLee/Xcode-Disk-Cleanup-Agent-Skill
xcode-project-analyzer	Auditing Xcode project config, build settings, schemes, and script phases for build-time improvements.	AvdLee/Xcode-Build-Optimization-Agent-Skill
xcsym	Symbolicating/triaging iOS/macOS crash reports (.ips, MetricKit, .crash, .xccrashpoint) — no other skill in this profile covers post-mortem crash analysis.	wraps CharlesWiltgen/Axiom’s tools/xcsym CLI (MIT); see docs/references.md for why only this tool was pulled in, not the full Axiom plugin

Notes on vendored skills

• swift-concurrency is also based on Antoine van der Lee’s paid Swift Concurrency Course — the skill body itself only carries a plain-text attribution line, no link: commit a228d797 deliberately scrubbed swiftconcurrencycourse.com/UTM links from the synced skill to avoid shipping course marketing fleet-wide. That decision stands here too — see the git history rather than expecting a link inside the skill.
• helm-asc documents the third-party Helm app’s helm-asc CLI, which is paid and not bundled. It was copied verbatim from the already-installed, already-working local skill rather than reconstructed from vendor documentation, so it describes what the operator’s licensed copy actually does.
• swiftui-liquid-glass additionally cites Apple’s WWDC25 session 323 for one design note.
• A rocketsim skill used to ship here and no longer does. It was consolidated onto FlowDeck along with the rest of this profile’s simulator/build-verification surface — see Build & test tooling (flowdeck).

supabase

Skill	Fires when	Source
supabase-postgres-best-practices	Writing, reviewing, or optimizing Postgres queries, schema, indexing, connection pooling, RLS policies, or locking.	supabase/agent-skills (official, MIT)

Third-party skills

Not this repo’s own skills — installed per-project via [project].skills in .lacquer.toml and lacquer skills (see the README for the mechanism). Listed here so the packages this fleet actually pulls in have a source link.

dpearson2699/swift-ios-skills (PolyForm Perimeter 1.0.0, not MIT/permissive — installed and used as distributed via skills add, not copied into this repo, which is what keeps this within the license’s terms) — Apple-framework reference skills. lacquer init suggests these automatically from a component’s Swift imports (see internal/skillsuggest); this table is that same mapping.

Skill	Swift import(s)
activitykit	ActivityKit
app-intents	AppIntents
apple-on-device-ai	FoundationModels
authentication	AuthenticationServices, LocalAuthentication
avkit	AVKit
background-processing	BackgroundTasks
cloudkit	CloudKit
cryptokit	CryptoKit
device-integrity	DeviceCheck
healthkit	HealthKit
mapkit	MapKit
musickit	MusicKit
pdfkit	PDFKit
photokit	PhotosUI
push-notifications	UserNotifications
realitykit	RealityKit
storekit	StoreKit
swift-charts	Charts
swift-testing	Testing
swiftdata	SwiftData
vision-framework	Vision, VisionKit
widgetkit	WidgetKit

HunterHillegas/mac-assed-mac-app-skill — AppKit/macOS app conventions. Not import-suggested (no single Swift import signals a macOS app); declared explicitly where needed (e.g. foxtrot).