Explorar skills

Branch-aware audit orchestrator. On main — scans whole project, opens tracker issue. On feature branch — scopes to diff, auto-fixes, stages changes.

Create distinctive, production-grade frontend interfaces. Rejects generic AI aesthetics (purple gradients, Inter font, glassmorphism). Design direction first, then implementation.

Read, extract and summarize PDF files with automatic OCR fallback for scanned/image-based PDFs. Use this skill whenever the user asks to read/extract/summarize PDF content. Triggers on: "read this PDF", "extract PDF content", "summarize this PDF", "what's in this PDF", "convert PDF to markdown", "batch process PDFs", "PDF 内容", "读取 PDF", "总结 PDF". This skill handles everything automatically — conve

专门用于深度分析和解读本地或 GitHub 源代码仓库的 Agent Skill。在 Claude Code 中全自动运行，包含质量 Review、迭代优化闭环、并行处理和 subAgent 支持。当用户说"帮我分析这个 GitHub 仓库"或"分析当前目录的代码"时，使用此技能。

Bootstrap a new "<vendor>-api-skill" project — a safe, cross-platform single-file Python CLI that mediates a third-party REST API, plus a matching Claude skill that drives it under explicit AI-safety constraints (token never enters AI context; every mutation passes through a six-tier safety classifier; every mutation is recorded in a local audit log). Trigger when the user asks to "build a CLI / s

Finds existing tools, libraries, MCPs, templates, datasets, and frameworks before creating anything from scratch. Use when the user asks to build, create, add, implement, write, draft, design, set up, automate, or analyze anything - especially on phrases like "build me a", "I want to create", "add X to my project", "is there a library for", "is there a template for", "how do I implement", "set up

Generate a customized Switch interview script based on your product context and existing evidence gaps. Outputs a ready-to-use markdown guide with Moesta's timeline reconstruction questions and force-probing techniques. Use when: "interview script", "jtbd interview", "interview guide", "how to interview".

Find cross-interview patterns across 3+ switch analyses in .jtbd/switches/. Clusters recurring jobs, identifies force patterns, surfaces evidence gaps, and generates actionable recommendations. Outputs structured YAML to .jtbd/patterns/. Use when: "find patterns", "jtbd patterns", "cross-interview", "what are the jobs".

Merge current PR after CI passes. Waits for all checks, squash merges, deletes branch, switches to main.

Delegates knowledge bank lookups to Explore subagents with reflections-first strategy that learns from past mistakes. Checks /reflections/ directory before documentation to extract documented failures to avoid and proven approaches to apply. Automatically triggers when user mentions services ([project-a], [project-b], [project-c], Claude Code), requests investigations, or queries past lessons. Ret

Tool-tier guidance for symbol search and call/reference tracing. Preloaded into agents that locate code; not user-facing.

Run code review with dynamic agent scaling (3-13 agents based on change size)