Top Agent Skills

Anti-slop frontend design skill for building NEW landing pages, portfolios, and UI from a brief. Reads the brief and existing brand context first, infers a design direction from three configurable dials (variance, motion, density), and ships interfaces that don't look templated or default-AI. Use when starting a new page/UI with no existing identity to preserve, or when asked to design something "that doesn't look like every other AI landing page." For auditing and upgrading an EXISTING UI, use `taste-redesign` instead — this skill is for the moment before code exists. Triggers on: "build a landing page", "design a portfolio", "make this not look generic/templated/AI-generated", "anti-slop", starting UI work with a brief but no established identity yet. Also owns component-library selection (which package for toasts, dropdowns, charts, drag-and-drop, state, etc.) — use this instead of hand-rolling a component or guessing at a dependency. Source: github.com/Leonxlnx/taste-skill (taste-skill, the main v2 varian

Audits an EXISTING UI/codebase for generic AI patterns and applies targeted craft fixes — layout, interactivity, content, iconography, and code quality — without discarding the project's own identity. Use when a UI looks flat, generic, or "AI-slop" and there is code to improve, not build from scratch. Triggers on: "improve the design", "looks generic", "not polished enough", "redesign this", "elevate the UI", "it looks boring", "make it better looking", "apply taste", "design review". Works with any CSS framework. For designing a NEW project's identity from a brief, use `taste-skill` instead — this skill only upgrades what's already built. Source: github.com/Leonxlnx/taste-skill (redesign-skill), restored to the full checklist + the project-identity rule the earlier port had dropped.

Encodes Emil Kowalski's design engineering philosophy — UI polish, component design, animation decisions, and the invisible details that make software feel great. Use as the reference/philosophy layer for interaction craft. For the narrower, task-specific pieces of this same philosophy, see `animate` (build one animation), `review-animations` (review a diff), `improve-animations` (audit a codebase), `find-animation-opportunities` (find what's missing motion), `animation-vocabulary` (name an effect), `apple-design` (gesture/spring physics), `ask-sonner` (the Sonner toast library). Source: github.com/emilkowalski/skills.

Translates stakeholder feature requests into validated outcome statements before any work is committed. Use this skill — proactively and without waiting to be asked — whenever a stakeholder, exec, or customer arrives with a pre-packaged solution: "we need a dashboard", "add a Slack notification", "build an export feature", "create a report", "let's add a filter", "can we just add X". Also triggers for: "how do I push back on this request", "what outcome does this feature solve", "outcome vs output", "outcomes not features", "what are we really trying to achieve", "we're being a feature factory", "I need to reframe this as a problem", "the stakeholder is pushing a specific solution", "discovery before delivery", "assumption testing", "translate this request into an outcome", "ship outcomes not features". Runs the 'One Framework. Four Questions.' protocol (Liatti + Cagan + Torres): Behavior Change → Assumption Test → Cheapest Test → Success Metric. Produces an Outcome Brief with embedded AI prompts ready to pas

Transforms technical git commits into a clear, compact, user-friendly changelog. Use when preparing release notes, creating product update summaries, documenting changes for customers, or maintaining a public changelog page.

Writes LinkedIn posts in the user's own voice, from a real observation or earned insight — not engagement-bait or growth-hacking templates. Use when drafting a LinkedIn post, turning a work observation or lesson into a post, or reviewing a LinkedIn draft for tone. Normally invoked by `writing-system` once LinkedIn is the chosen platform, but works standalone too. Companion skill: `x-thread-writer`, for the same voice on X/Twitter.

Self-contained code review checklist — formatting, type safety, duplication, readable conditions, logic-out-of-UI, test coverage, comment quality. Adapts to whatever tooling the project actually has (skips Prettier/TypeScript/test-runner checks if unconfigured). Use before merging, after completing a feature, or any time you want a second pass on code quality. In Claude Code, prefer the official `code-review` skill or `pr-review-toolkit`/`feature-dev` code-reviewer agents instead — both are real, installed, and go deeper than this checklist. This copy exists for Codex/Cursor portability, where those agents aren't available.

Structured verbal communication framework for Senior PM and TPM interviews. Use when preparing for interviews, rehearsing answers, practicing English (or Spanish) delivery under pressure, or when any interview question needs to be answered in a crisp, structured way. Covers SCQA/STAR/C-F-I scaffolding, answer templates by question type, pushback handling, English- and Spanish-under-pressure tactics, and post-interview debrief methodology. Trigger on any variant of "how do I answer this?", "rehearse with me", "how would you say this?", or "practice interview question".

Expert jq usage for JSON querying, filtering, transformation, and pipeline integration. Practical patterns for real shell workflows.

Database management, forms, reports, and data operations with LibreOffice Base.

Evaluate and rank agent results by metric or LLM judge for an AgentHub session. Use when the user runs /hub:eval or asks to score, compare, or pick a winner among completed AgentHub agents.

Build content-focused websites with Astro — zero JS by default, islands architecture, multi-framework components, and Markdown/MDX support.

Documentation and capabilities reference for Daily

Set up a new autoresearch experiment interactively. Collects domain, target file, eval command, metric, direction, and evaluator. Use when the user runs /ar:setup or asks to start optimizing a file with the autoresearch loop.

Freeze

Interact with GitHub issues - create, list, and view issues.

Fast in-memory DataFrame library for datasets that fit in RAM. Use when pandas is too slow but data still fits in memory. Lazy evaluation, parallel execution, Apache Arrow backend. Best for 1-100GB datasets, ETL pipelines, faster pandas replacement. For larger-than-RAM data use dask or vaex.

Analyze auto-memory for promotion candidates, stale entries, consolidation opportunities, and health metrics. Use when the user runs /si:review or asks what has been learned and what should be promoted or pruned.

Memory health dashboard showing line counts, topic files, capacity, stale entries, and recommendations. Use when the user runs /si:status or asks how full or healthy the agent memory is.

Document creation, format conversion (ODT/DOCX/PDF), mail merge, and automation with LibreOffice Writer.

Expert C4 Code-level documentation specialist. Analyzes code directories to create comprehensive C4 code-level documentation including function signatures, arguments, dependencies, and code structure.

GraphQL gives clients exactly the data they need - no more, no less. One endpoint, typed schema, introspection. But the flexibility that makes it powerful also makes it dangerous. Without proper controls, clients can craft queries that bring down your server.

Inngest expert for serverless-first background jobs, event-driven workflows, and durable execution without managing queues or workers.

Use mmx to generate text, images, video, speech, and music via the MiniMax AI platform. Use when the user wants to create media content, chat with MiniMax models, perform web search, or manage MiniMax API resources from the terminal.

Safe Notes Skill

Draft documentation updates with governed file-write proposals.

Draft a governed social content calendar without posting externally.

Inspect package metadata, lockfiles, scripts, and npm release readiness.

Review docs, tool output, skills, and memory candidates for prompt-injection risk.

Create OR review a wireflow — the artifact BETWEEN user journeys and wireframes. It maps whole journeys into swimlane flows with high-level navigation (screens, decisions, system/agent steps) while keeping Jobs-To-Be-Done at the core, WITHOUT deep UI. Use whenever the user wants to "map the flows", "make/build a wireflow", turn journeys / JTBDs / a spec / a live prototype into flows, or put every journey on one board against shared owner lanes — even if they never say "wireflow". ALSO use it to REVIEW or critique an existing wireflow (image, FigJam, or description). In the A-Team pipeline this is a definition-phase skill: output lands in docs/features/<slug>/briefs/wireflow/ with jobs consumed by id from docs/product/jtbd/; pipeline mode derives the method decisions and highlights the riskiest at the gate, standalone mode runs the full grill. CREATE generates verified SVG/HTML (self-checked by rendering and Reading its output) in a horizontal per-journey OR shared-matrix layout, and can rebuild in FigJam. Do

ateam-design

ateam-discovery

product-brainstorming

Load this skill when the user says "plan milestone", "plan-milestone", or "/plan-milestone". Derives sub-issues for a milestone epic, persists all artifacts to disk, enforces UI enrichment via Stitch, and creates GitHub issues only after full approval.

Load this skill when the user says "verify-implementation", "verify implementation", or "/verify-implementation". Reviews the implementation against the issue requirements and project coding standards, runs automated checks, and produces a gated report before acceptance criteria are ticked off.

Write idiomatic QML for Qt Quick applications with sound property design, bindings, and C++ integration.

Fan one question out across every connected source and return a single ranked, deduplicated, attributed result set.

Turn research into a decision-ready brief tailored to a specific stakeholder and decision.

Onboard a new clinic student onto a matter and the clinic's systems fast, with supervision and confidentiality from day one.

Write and review interface microcopy - errors, empty states, buttons, labels, and CTAs - that is clear, concise, human, and consistent with the product's voice.

Write QDoc API documentation for Qt6 C++ code, covering classes, properties, signals/slots, enums, and QML-exposed types.

Repurpose one piece of content into platform-native formats across short-form, social, and email without copy-pasting.

Profile an unfamiliar dataset — shape, grain, quality, nulls, distributions, and duplicates — before any analysis is trusted.

Write a project charter that pins down goal, scope, success criteria, and stakeholders before work starts.

Write a newsletter issue built around one core idea, with a subject line, tight structure, and a single CTA.

QA a completed analysis for methodology, accuracy, and bias before it is shared, shipped, or acted on.

Write or maintain technical documentation — READMEs, runbooks, API references, architecture docs — that a specific reader can act on without asking follow-ups.

Adapt content for a specific locale — units, dates, currency, idioms, and cultural fit — beyond literal translation.

Tailor the IP-LEGAL sub-plugin to one organization by capturing its risk tolerance, jurisdictions, and default playbook positions.

Turn a raw transcript or scribble into decisions, action items with owners and dates, and open questions - so nothing agreed in the room evaporates afterwards.

Maintain a RAID log tracking risks, assumptions, issues, and dependencies with owners and review cadence.

Run a structured, multi-source deep-research loop on a question until the answer stops changing.

Adapt the commercial-legal playbook and its standard / fallback / walk-away positions to a company's risk tolerance, industry, and signing authority, and version the result.

Trace a break between two records of the same position or cash — accounting vs custodian, fund vs prime broker, trade vs settlement — from the aggregate difference down to the offending transaction, and classify it for clearing.

Organize symptoms, history, and questions into a clear summary to bring to a clinician — no diagnosis.

Document architectural decisions in standardized ADR format. This skill should be used when making significant technical decisions, choosing between approaches, establishing patterns, or when other skills (create-plan, implement-plan, brainstorm) identify decisions that need documentation. Triggers on "document decision", "create ADR", "architectural decision", or automatically when invoked by other skills during planning and implementation.

Full autonomous engineering workflow

Google quantum computing framework. Use when targeting Google Quantum AI hardware, designing noise-aware circuits, or running quantum characterization experiments. Best for Google hardware, noise modeling, and low-level circuit design. For IBM hardware use qiskit; for quantum ML with autodiff use pennylane; for physics simulations use qutip.

Use this agent when you need to run linting and code quality checks on Ruby and ERB files. Run before pushing to origin.

Rice