Find the open-source solution that fits.

At the end of a meaningful conversation, the skill scans the full session for durable decisions, preferences, pitfalls, and open questions rather than producing a generic transcript summary. It presents a short list for editing and writes only after explicit confirmation, using KnowMine, another MCP-compatible notes tool, or clean Markdown for manual storage.

Describe a workload such as daily chatbot volume or document processing, and the skill applies provider pricing, token formulas, caching, and batch discounts to estimate the bill. It compares models across OpenAI, Anthropic, Google, DeepSeek, Mistral, and others while showing the math. Because model prices change frequently, its reference files need regular updates before budgeting decisions.

For complex architecture, review, or security tasks, the skill dispatches targeted prompts to Codex CLI and Gemini CLI at the same time and returns structured outputs for Claude to synthesize. It highlights agreement and divergence instead of treating either model as authoritative. Claude Code, both external CLIs, and Python 3.8 or newer must already be installed and authenticated.

Provide a brand name, product idea, audience, or a few constraints, and the skill develops three distinct visual directions, selects a recommendation, and writes a strategy document plus a standalone HTML brand board. It can also generate CSS design tokens for handoff. Invented testimonials, logos, or metrics must be labeled as concept content rather than presented as real proof.

Define highlights, numbered callouts, labels, steps, and arrows against Playwright locators or selector strings. The tool injects those annotations as DOM elements before capture, so they render sharply and stay anchored to the interface. Replayable JSON specs let documentation screenshots be regenerated after UI changes. It requires Node 18 or newer, Playwright, and an accessible web page.

Provide a topic, an optional research file, and an iteration count. Each fresh subagent reads the document's findings, investigated paths, decision log, and open questions, then spends its rounds only on unresolved gaps before updating the same file. Dead ends are marked `UNFILLABLE`, and the document is the sole handoff between iterations, preventing later passes from rediscovering already covered facts.

This Claude Code skill turns package targets, CVE advisories, Jira tickets, or Dependabot alerts into an upgrade workflow for Python, JavaScript/TypeScript, and Go. It analyzes direct and transitive dependencies, reviews breaking changes, proposes confirmed code edits, runs layered tests, and can finish with commits, pull requests, and Jira updates.

This skill uses Pencil MCP to translate design files into UI code for iOS, Android, Flutter, or web projects. Its staged workflow saves progress checkpoints, can run with parallel agents or a slower linear mode, and uses differential audits to update only the parts affected when a design changes.

This bilingual Claude Code skill applies Sun Tzu-inspired frameworks to real business and technology decisions rather than returning quotations. It questions the situation, classifies the competitive terrain, examines asymmetries, selects a strategic approach, and turns the analysis into concrete recommendations for positioning, market entry, negotiation, or team conflicts.

session-tab-namer gives every new Claude Code session an immediate fallback label, then replaces it with a semantic name such as `build:*`, `research:*`, or `debug:*` once the objective is clear. It works across common terminal emulators and tmux, persists names for other tools, and uses a SessionStart hook whose installer requires jq.

second-opinion intercepts plans involving migrations, authentication, billing, infrastructure, production operations, or broad file changes before execution. It routes them to Codex, Gemini, or Opencode for a compact SHIP, REVISE, or RECONSIDER verdict, and falls back to a fresh-context self-review when no external reviewer CLI is available.

This Claude Code skill guides Korean users through personal-finance questions covering take-home income, social insurance, spending, savings, debt, taxes, housing, pensions, insurance, and investing. Its reference material is organized around 2026 Korean rules such as tax rates, DSR limits, housing subscriptions, and retirement programs, producing structured plans from a user's situation.

Zilliz Launchpad guides an agent through collecting data, configuring a use case, planning an index, ingesting records, evaluating search, and deploying a runnable application. It supports text and multimodal inputs through an agent skill or MCP server, targeting local Docker-based Milvus or Zilliz Cloud, with Python, Node.js, an embedding API key, and the chosen deployment environment required.

This Claude Code skill turns product details and one of three visual directions—technology, Chinese-inspired, or warm—into concise Chinese generation prompts for Doubao and Jimeng. It produces text-free and branded logo variants, invitation key visuals, short slogans, and an iteration checklist so the same visual concept can carry into video and other launch assets.

This client-specific skill coordinates five agents for topic selection, outlining, article writing, imagery, layout, and delivery of NSKSD nattokinase distributor content. It adds 30-day topic deduplication, compliance and formatting gates, Feishu multi-select callbacks, and a fallback to Feishu documents when the WeChat draft box cannot be used.

Use this Claude skill to turn deal details into a customized Customer Agreement based on Y Combinator’s standard SaaS form. It applies predefined drafting decisions, removes template annotations, and produces both an editable DOCX and a memo explaining the changes. The result is a starting point, with issues such as data processing and IP ownership explicitly flagged for counsel review.

This agent skill guides users through France’s income-tax return, including Form 2042 and common annexes. It maps salary, business, property, investment, crypto, and foreign income to specific boxes, builds a document checklist, and can estimate tax while comparing PFU with the progressive scale. It does not file the return, and its projected 2026 figures should be checked against official guidance or a tax professional.

Give the skill your LinkedIn profile PDF, photo, and examples you admire, and it will score the profile section by section, rewrite weak copy, and create a niche-specific editorial calendar. It can also prepare posts with PNG images or carousels and adapt future batches from engagement data you provide. Publishing remains manual by default; image generation and scheduled posting require optional Gemini or n8n setup.