Browse skills
NVIDIA FastGen Runtime for Claude Code: Verify Paths and Dry-Run Settings
FastGen Runtime resolves the working directory, Python environment, config path, output paths, and run identity before a FastGen command is executed.
nvidia/skillsClaude CodeFastGen Setup for Codex: Verify the Local Environment
Check a local FastGen environment before training. Use this NVIDIA skill to trace the active checkout, dependencies and required access for the intended task.
nvidia/skillsCodexFastGen Train for Claude Code: Plan a Bounded Smoke Test
Use FastGen Train to prepare a bounded NVIDIA FastGen trial, verify checkpoint assumptions and define which optimizer phases the test should exercise.
nvidia/skillsClaude CodeNVIDIA FastGen Validate for Claude Code: Plan a Reproducible Parity Check
FastGen Validate describes how to preserve comparison settings, locate the first divergent intermediate, and report numerical or visual evidence separately.
nvidia/skillsClaude CodeFastHTML and pywebview Desktop App Skill for Claude Code
Plan and review a Python desktop app with a local FastHTML service, HTMX screens and a pywebview window. Includes references for packaging.
kuailexiaozixin/fasthtml-desktopClaude Code / Codex / DeepSeek HarnessClaude Code Feature Flags Skill for React Channel Tests
Trace a React flag across the shared configuration and channel forks, then decide how to test the feature. This is useful when a component test behaves differently on canary and www.
facebook/reactClaude CodeCheck scope and affected files before implementation in Claude Code
Restate the feature goal, inspect related code, and present affected files, tradeoffs, and open assumptions for confirmation.
snoodleboot-io/tideraceClaude CodeClaude Code feature-prioritizer skill for ranking features with RICE
Rank comparable quarterly estimates by reach, impact, confidence, and effort. Mark committed work separately.
hamzafarooq/claude-code-starterClaude CodeChatGPT fec-backend-requirements-handoff skill for UI requirements
Turn screens and user actions into a Markdown handoff covering data needs, UI states, business rules, and open questions.
bovinphang/frontend-craftChatGPTFeedback Analyzer Skill for ChatGPT: Cluster Comments and Keep Evidence
Label customer comments by topic and sentiment, link each group to input evidence, and record the sample size. Use the clusters to decide what merits follow-up.
inbharatai/claude-skillsChatGPTAI Feedback Loops Skill for ChatGPT: Design Partial Edits and Undo
Turn ratings, regeneration, inline editing, and direct corrections into observable feedback. Specify timing, user actions, system response, and how accepted content stays in place.
Owl-Listener/ai-design-skillsChatGPTClaude Code SBI Feedback Skill for Concrete Workplace Examples
Prepare feedback around a specific situation, an observable behavior and its actual impact, with room for the other person's account.
davila7/claude-code-templatesClaude CodeFetch Zulip Messages for Claude Code: Read a Web-Public Topic Link
This Zulip contributor skill retrieves context for a narrow topic link using spectator access, so a developer can read the relevant public discussion from a project issue.
zulip/zulipClaude CodeCodex DbSqlLikeMem Fidelity Tests for Mock and Provider Parity
Compare a mock and a real local provider on observable behavior such as parameter handling, result shape, and failure contracts.
christianulson/DbSqlLikeMemCodexHyperFrames Figma Import Skill for Codex Video Projects
Import a Figma asset or frame as local footage or editable source. Preserve design colors and place the imported component on a HyperFrames timeline.
heygen-com/hyperframesCodexCodex Figma Component Audit for Missing Button Variants
Audit a Figma library inventory for naming consistency, state coverage and prioritized fixes.
spontus-reach/spontusCodexCodex Figma design briefs and screen acceptance
Translate a product request into goals, flows, screens, reusable components, and observable edge-state requirements.
spontus-reach/spontusCodexCodex Figma Product Design Review for Empty-State Coverage
Compare a Figma design with its product requirements and user flow, then report blockers, open questions and a proposed approval status.
spontus-reach/spontusCodexFile Organizer Skill for Dry-Run Plans in Claude Code
Inventory a scoped folder, group duplicate content, and produce a reviewable move plan with path collisions identified.
davila7/claude-code-templatesClaude CodeFind a Dropbox File by Name in a Test Folder With Claude Code
Search an authorized Dropbox folder for a known document and report a short set of candidates with paths that distinguish similar file names.
openai/pluginsClaude CodeCodex QAIO Help Center Gap Ranking from Local Tickets
Compare repeat questions with existing help articles, then rank uncovered topics by frequency, freshness and available customer weights.
akiotanaka847/qaio-desktopCodexFind Skills for Claude Code: Search by Task and Review GitHub Sources
Find Skills turns a concrete task into a skills search and helps compare results by owner, repository, and relevance before installing anything.
vercel-labs/skillsClaude CodeFind Style Issues Skill for ChatGPT: Triage Long Paragraphs, Headings and Shell Lines
Review supplied documentation findings against project thresholds. Triage long prose and shell lines, flag heading candidates, and incorporate an existing Vale report when available.
giantswarm/docsChatGPTClaude Code React Fix Skill for Formatting and Lint Checks
Use the React repository's formatter and lint tasks on current changes. Review what the formatter handled and locate any lint findings that still need work.
facebook/reactClaude CodeCodex DbSqlLikeMem Fix Build for .NET Diagnostics and CS1591
Trace the first blocking build diagnostic to its source file, reference, or target framework. Apply a focused fix and rerun the same build.
christianulson/DbSqlLikeMemCodexMarkdown Lint Fix Skill for ChatGPT: Repair Code Fences, Tables and Line Length
Use supplied lint findings to locate specific lines. Draft minimal fixes for fence languages, table separators, long prose, and formatting while preserving URLs and inline code.
multigres/multigresChatGPTCodex DbSqlLikeMem Fix Tests for SQL Provider Behavior
Reproduce failing tests and compare mock results with the real database provider. Keep regression coverage and preserve genuine provider-specific behavior.
christianulson/DbSqlLikeMemCodexCompare React Release-Channel Feature Flags With Claude Code
Read the feature flag report for the checked-out React commit and compare exact values across release channels.
facebook/reactClaude CodeDiagnose Flaky UI Tests With Codex
Compare repeated results and artifacts to distinguish selectors, timing, shared state and environment failures.
anhtester/codex-testing-kitCodexClaude Code React Flow Skill for Renderer-Specific Type Checks
Select the Flow target for DOM, Native, or Fabric in a React source checkout. Use it to locate type errors and verify constraints for the renderer affected by a change.
facebook/reactClaude Code