Explorar skills

Enforce a strict Gitflow-based workflow with conventional commits, semantic versioning, and issue-driven branching. Use when the user asks to commit, create a branch, open a PR, tag a release, or perform any git operation. Also applies when mentions 'commit', 'branch', 'merge', 'release', 'hotfix', 'gitflow', 'conventional commit', 'semantic versioning', or 'semver'.

Convert Markdown files to professionally styled PDFs. Use when the user wants to export a .md file as PDF for sharing with clients or external stakeholders. Supports tables, code blocks, CJK (Chinese) characters, multiple templates, and batch conversion.

Generates highly optimized Agent Skills for both native Claude Code and the Antigravity system, according to official best practices and the "Progressive Disclosure" strategy. Use when the user wants to build a new skill, turn a workflow into a skill, upgrade an existing skill to latest best practices, audit a skill for compliance, or structure agent instructions for either ecosystem.

Diagnose spreadsheet structure and recommend practical Excel/WPS-compatible formulas and workflows for cleaning, lookup and matching, summarization, text and date cleanup, and common formula troubleshooting. Use when the user asks what formula to use, how to process a table, how to clean messy columns, how to match sheets, how to make a solution work in both Excel and WPS, or when headers and samp

Use when investigating a topic that benefits from a reusable, citation-backed NotebookLM corpus—literature reviews, due diligence, regulatory/legal research, competitive intelligence, long-term personal knowledge queries, or any repeated deep-dive across the same body of material. Not for one-off web search, single-document reads, or code exploration.

Build and maintain a persistent knowledge base — a compounding vault of interlinked markdown files. TRIGGER when: user adds sources to raw/ (articles, papers, notes, images), pastes content to ingest, says 'add to wiki' or 'ingest this', asks synthesis questions ('what do we know about X', 'compare A and B', 'summarize research on Y'), requests health checks ('lint the wiki', 'find gaps'), or runs

Evaluate whether rewriting the user's prompt for clarity, typo correction, or actionability would meaningfully improve it; rewrite only when it would. Use when the user says "rewrite my prompt," "fix the typos in this prompt," "make this prompt clearer," "make this more actionable," or when a prompt is short and the intent is ambiguous. Display the rewrite and wait for approval.

🧠 Enable AI coding agents to adopt product management skills and build user-focused software efficiently.

Delegate coding tasks to Codex CLI for execution, or discuss implementation approaches with it. CodeX is a cost-effective, strong coder — great for batch refactoring, code generation, multi-file changes, test writing, and multi-turn implementation tasks. Use when the plan is clear and needs hands-on coding. Claude handles architecture, strategy, copywriting, and ambiguous problems better.

FlowMem generates a Mermaid-based AI memory system for any codebase, library, or documentation set. Every piece of knowledge becomes a plain .mmd file — human-readable in VS Code with the Mermaid Preview extension, renderable in any browser via the generated HTML viewer, traversable by any AI coding assistant. TRIGGER when the user says any of: - "generate flowmem", "build flowmem memory", "create

前端设计工作室。识别工作区是否有 DESIGN.md（Google Labs 2026/04 开源的视觉契约规范）：有就严格按它工作；没有且用户明说要"为这个项目制定视觉风格"时进入风格制定流程，多个变体的可交互预览页 → 多轮迭代 → 写出符合 9-section 规范的 DESIGN.md；其他前端任务下温和提示但不打断。所有前端工作都先在浏览器里给用户可交互预览页确认、不让用户读代码。触发场景：用户要做 landing page、仪表盘、原型、组件库、移动端 mockup、交互原型；说"做个页面/组件/界面/UI/前端"；说"为这个项目定个视觉风格/做一套视觉系统"。

Use this skill to query data with Cassis (which grounds the question against the user's project ontology and returns SQL, a plan to approve, or an ontology gap to clarify) and Snowflake (which executes the SQL). Triggers when the user says "ask cassis", "query with cassis", "cassis question", or asks any natural-language data question while the Cassis MCP and a Snowflake MCP are both connected. Ha