39 skills · sorted by relevance
Extend the amount of text a language model can process at once. Compares RoPE, YaRN, ALiBi, and position interpolation, with implementation examples, additional training guidance, and memory optimization approaches for long document processing.
Helps extend how much text a language model can process at once. Covers choosing RoPE, YaRN, ALiBi, or position interpolation, implementing position information, fine-tuning on longer documents, and managing memory during inference.
Manage Node.js dependencies with pnpm, from installing packages to coordinating multiple projects in one workspace. The skill covers shared version settings, third-party package patches, dependency overrides, and consistent installation procedures for automated build environments.
Build branching dialogue and narrative — a node/choice graph with conditions, variables, and localization hooks — and choose between authoring tools Ink and Yarn Spinner or a custom data-driven runner. Engine-neutral. Use when the user mentions dialogue system, branching dialogue, conversation tree, choices, Ink (.ink), Yarn Spinner (.yarn), or NPC dialogue.
The English summary is being prepared. Showing the original description.
Review, improve, rewrite, or draft Storybook documentation in /docs. Assess each page’s purpose, audience, structure, and examples, then make suitable edits or provide a content plan, with formatting checks after changes.
Prepare an existing Twenty app for npm or marketplace publication by reviewing its README, listing details, package metadata, logos, and screenshots. The workflow includes checking publication settings, validating the build, and running the publishing command.
Estenda janelas de contexto de modelos transformer usando RoPE, YaRN, ALiBi e técnicas de interpolação de posição. Use ao processar documentos longos (32k-128k+ tokens), estender modelos pré-treinados além dos limites de contexto originais ou implementar codificações posicionais eficientes. Cobre embeddings rotativos, vieses de atenção, métodos de interpolação e estratégias de extrapolação para LLMs.
The English summary is being prepared. Showing the original description.
Knip (TypeScript / JavaScript 未使用コード検出ツール) リファレンス。 未使用ファイル・export・依存関係・型・列挙メンバー・循環依存 (circular dependencies / cycles) 検出、 auto-fix、monorepo (workspaces)、catalog (pnpm / Yarn / Bun catalogs)、 knip.json / knip.config.ts、プラグイン、ignore、reporter (SARIF)、 初回クリーンアップ (first cleanup)、段階導入 (adopt gradually)、how Knip works。
The English summary is being prepared. Showing the original description.
Set up automated checks before each Git commit using Husky and lint-staged. Format staged files with Prettier, run existing type checks and tests, and verify the setup using the repository's package manager.
Organize software releases on GitHub, from version suggestions and release notes to testing, builds, and deployment. Coordinate multiple packages or repositories, roll changes out in stages, and configure rollback to restore an earlier release when problems occur.
Check out a GitHub pull request, fix linting and TypeScript errors, and commit and push the changes. The workflow supports forked pull requests and repeats type checks to verify fixes while keeping changes focused on these errors.
Rebuild modified packages in the Storybook monorepo and restart its internal UI after code changes. The workflow provides the running URL and can optionally create a UI review of stories related to the changes.
Update all Storybook packages in a project to a chosen release or canary, an early development build. Use it to test changes in external apps, reproduce bugs, and keep package versions aligned while handling peer dependencies.
Plan and carry out major library and framework upgrades by checking compatibility, identifying changes that require code updates, testing each stage, and preparing rollback steps if an upgrade causes problems.
Review MCP configurations for exposed credentials, risky command patterns, and package versions that can change unexpectedly. Produce a report with severity levels and suggested fixes without running the commands defined in the configuration.
Review websites for security, browser compatibility, and code quality. Identify issues in HTTPS settings, third-party libraries, HTML, and error handling, then apply concrete improvements using checklists and implementation examples.
Design and review system architecture, compare databases and structural approaches, and document technical decisions. Included tools generate architecture diagrams, inspect dependencies, and assess project structure to identify circular relationships, excessive coupling, and opportunities for improvement.
Create and run tests for interactive command-line tools and terminal interfaces using @microsoft/tui-test. Simulate keyboard input, check visible text, and compare screen snapshots in a separate terminal environment for each test.
Audit npm, PyPI, and Go dependencies for version-specific advisories, inactive upstream projects, concentrated npm publishing rights, and installation scripts. Produce a report backed by collected data, with assessment coverage, action priorities, and upgrade or replacement guidance.
Build remote MCP servers, which let AI agents call external tools, on Cloudflare Workers. This skill covers server code, tool definitions, OAuth authentication, local testing, deployment, and connecting a Codex client.
Build and deploy AI agents on Cloudflare Workers using the Agents SDK. Supports persistent conversation state, live WebSocket connections, scheduled tasks, and tool integration, with guidance for connecting browser or React chat interfaces.
Review web code for security, browser compatibility, and code quality. Covers HTTPS, dependency vulnerabilities, security headers, outdated APIs, HTML structure, error handling, and cleanup of event listeners, with examples and audit checklists to guide improvements.
Check release readiness by reviewing tests, type checks, linting, builds, security findings, documentation, and environment settings. Get a report separating deployment blockers from warnings, plus action items and checks for post-deployment health and rollback preparation.
Unauthorized access playbook for common exposed services. Use when Redis, Rsync, PHP-FPM, AJP/Ghostcat, Hadoop YARN, H2 Console, or similar management interfaces are exposed without authentication.
The English summary is being prepared. Showing the original description.