Shape agent capabilities into your workflow.

frontend-post formats user-supplied Chinese copy into Xiaohongshu-style cards at fixed 1080×1440 or 1080×1080 dimensions. It offers twelve curated visual presets and exposes colors, fonts, decoration, spacing, and copy limits for direct adjustment, while leaving the content itself under the user's control. PNG export uses Chrome with Puppeteer or Chromium with Playwright; it is a layout tool, not a copy generator.

Travel Agent turns a trip idea into a seven-phase plan: clarify priorities, shape destination options, check live logistics, build a paced day-by-day itinerary, track reservations, prepare packing, and export a local PDF. Each day gets a named anchor and backup, with advice shaped around the traveler's interests. It is not a booking or fare-aggregation service, and its source-available license is non-commercial.

shopping-compare researches purchases across Chinese e-commerce platforms, combining cross-site comparisons, independent reviews, safety records, and checks for misleading marketing claims. It starts by asking whether the purchase is necessary, then gives reasons, alternatives, and risks while leaving the final choice to the shopper. Direct access to several platforms is blocked, style products may require screenshots, and web prices can become outdated.

Bernard Stiegler Perspective Skill loads Stieglerian concepts and methods into Claude or another compatible agent for humanities research. Users can ask for a Stieglerian reading of material about AI, platforms, algorithms, or digital industry, then choose a default or lineage-specific angle. It is an analytical lens rather than a conversational character or general-purpose assistant.

sync-fork installs a GitHub Actions workflow that discovers a fork's upstream parent and runs gh repo sync on a recurring schedule, defaulting to every six hours. It performs no commits, rebases, or force-pushes; conflicts fail loudly instead of being hidden. A logged-in GitHub CLI is required, and upstream workflow changes may require a separate token with the workflow scope.

claude-memory-skill stores structured conversation summaries, decisions, rejected ideas, open questions, entities, and tags in a Supabase database. Use /mem to save and /context or /recall to search and restore a topic in a later conversation, device, or interface. It requires a configured Supabase project and credentials, so users must handle access control and the sensitivity of the material they choose to store.

google-chrome-skill documents macOS workflows for discovering Chrome profiles, editing bookmark JSON safely, controlling tabs and windows with AppleScript, and locating extensions. It also describes cookie extraction for authenticated internal API calls when an MCP lacks the needed action. The skill is macOS-only, and cookie handling should be treated as sensitive because it can expose logged-in browser state.

audio-to-action transcribes a recording, cleans speaker sections, classifies one of five scenes, and selects a matching report template for meetings, discussions, memos, or progress updates. Tasks and message drafts carry timestamp evidence and explicit uncertainty, while external sending always waits for user approval. The current version has no real speaker diarization, automatic splitting of very long audio, cross-meeting memory, streaming, or automatic sending, and needs a configured ASR endpoint; authentication, when required, comes from an environment variable.

clarify interrogates a request before work begins, identifying assumptions and presenting up to four structured, clickable questions with recommended options. It is useful when an incorrect implementation would cost more than a few extra messages, but it is manually invoked and never activates just because a prompt happens to match its description. The local plugin collects no data.

ai-phrase-check uses a Detect, Suggest, Apply flow to spot recurring AI-style phrases in English and German and offer more human alternatives. The Claude Code skill and its in-progress Next.js demo share the same Markdown phrase lists, with user gates between stages; browser detection runs locally, while LLM suggestions use a user-supplied key stored only in session storage. The public demo is not live yet and there is no backend.

TradingView Session Skill turns the startup ceremony into a single tv_start trigger that launches TradingView Desktop, checks the local MCP connection, verifies a rules.json knowledge base, and returns a compact readiness status. It has a fallback launch path when the normal MCP action fails. TradingView Desktop, a local companion MCP server, and a machine-specific rules file are required.

Marketing Audit Skill gathers audits, SEO analysis, copy and ad drafting, funnel reviews, competitive analysis, and launch planning behind one /marketing dashboard. Fifteen sub-skills use a shared scoring rubric and require recommendations to name the change, location, and expected impact. Lazy loading keeps each run focused, and missing pages or inputs are reported instead of filled with invented analysis.

restate-confirm makes Claude Code summarize its understanding in one or two declarative sentences before editing files, deleting data, running side-effecting commands, or acting on an ambiguous request. It waits for confirmation, repeats the understanding after a correction, and then proceeds. Clear continuation signals and ordinary questions can skip the protocol; the plugin is a local instruction layer rather than an execution guard.

clean-code-audit scans an existing codebase for long files and functions, complexity, weak comments, magic literals, oversized tests, and duplication, then writes a prioritized Markdown report without editing anything. Python analysis uses the AST, while TypeScript and JavaScript rely on a documented heuristic. Thresholds and scan scopes are configurable, and the findings are intended to guide a separate refactoring effort rather than trigger automatic changes.

Agency OS Roadmap App Skill turns SETUP-ROADMAP.md into a self-contained HTML tracker with phases, weekly timing, estimates, critical-path markers, resource links, checkboxes, and per-task notes. Progress stays in localStorage by default, while an optional Supabase mode synchronizes workspaces and falls back offline. The current skill and UI strings are primarily German, and team sync requires a configured Supabase project.

claude-skill-pptx-preview opens a PPTX in PowerPoint for Mac through AppleScript, exports a PDF, and converts each page to a PNG with pdftoppm for visual inspection. Using PowerPoint as the renderer makes font fallback and complex-shape behavior closer to what the user will see. It requires macOS, PowerPoint, poppler, and automation permission, and the output resolution is adjustable.

OWASP Mobile Audit Skill checks Flutter, Android, and iOS codebases against all ten OWASP Mobile Top 10 categories plus six emerging checks. The report groups findings by severity, includes file and line evidence with a concrete fix, and records positive controls already present. You can filter the run by rule IDs, but the Markdown report remains an audit aid rather than proof that a mobile app is secure.

Flutter Dependency Audit Skill reviews a Flutter or Dart project's packages across outdated versions, abandoned libraries, duplicate functionality, license conflicts, known CVEs, loose pinning, and dependency overrides. It reports package-level evidence, severity, positive findings, and prioritized next steps instead of changing the dependency graph automatically. A pubspec.yaml and ideally pubspec.lock are required for a useful audit.