Find the open-source solution that fits.

Import existing CLI and VS Code sessions into Claude Desktop by creating pointer records for transcripts the app does not list yet. The skill performs a dry run, flags live sessions, archives copies, skips duplicates, and supports undo. It is macOS-specific, requires Desktop Code mode to have opened once, and needs an app restart before imported sessions appear.

Transcribe speech locally with faster-whisper, type it into the focused Claude Code window, and submit when you say phrases such as “go” or “send it.” The listener supports pause, resume, editable vocabulary, and a learning toggle across macOS, Linux, and Windows. The first setup downloads a model; macOS also needs microphone and accessibility permission.

Use the skill to implement or debug Taiwan ezPay e-invoice calls for issuing, voiding, allowances, queries, carrier checks, and invoice-number tracks. It explains AES-256-CBC, the reversed CheckValue and CheckCode ordering, and separate store and member keys, with a Python verifier for offline vectors and an end-to-end smoke test. The documentation was checked against dated versions; confirm the official manuals before use.

Describe a dream and the skill separates observable details from waking-life context and personal associations before offering readings. It presents the best-supported interpretation first, labels depth-psychology angles as hypotheses, and ends with a question rather than a one-shot decoding. Its knowledge graph draws on 18 historical, clinical, and scientific sources, while treating symbols as context-dependent.

Use the generator for a new system: it interviews for brand preferences and writes a readable spec, visual HTML guide, and project reference. Use the importer when a system already exists in Figma; it extracts design data and syncs code tokens while keeping the guide aligned. Live Figma import needs the relevant MCP, though token exports or screenshots can also be used.

Run formatting, Clippy, dead-code, unused-dependency, RustSec, test, and production-code heuristic checks before opening a pull request. PREPR intersects findings with changed files, separates new blockers from pre-existing context, and reports a severity-ranked verdict. Missing analysis tools can be installed automatically unless disabled.

Keyword Intelligence provides vendor-neutral, offline keyword analysis from tool-exported CSV files. It normalizes and enriches terms, applies scope and scoring steps, identifies gaps, and produces five deliverables for English, French, German, or Spanish work. It makes no external API calls and needs no credentials; XLSX or JSON must be re-exported as CSV.

Apple Reviewer Simulator drives a booted iOS Simulator to install and exercise a real build. It checks crashes, placeholders, sign-in, permissions, account deletion, user-content reporting and blocking, purchases, and device UI details, then writes an Apple-style verdict with guideline references, reproduction steps, and severity. The user supplies the build and simulator.

This Claude Code integration exposes VulnCheck CLI capabilities for CVEs, package lookups by PURL, exploit code or proof-of-concept searches, detection rules, and local project scans. It can also create signed backup URLs. VulnCheck CLI must be installed and authenticated; Windows users need WSL or Git Bash.

Scrollfilm Skill helps build Apple-style landing pages where scrolling scrubs a video or frame sequence. It covers loading overlays, text fades, parallax, section alternation, dots, and progress indicators, plus FFmpeg frame extraction, desktop/mobile trade-offs, iOS limitations, and reduced-motion handling.

ProveIt asks an LLM to attach a direct URL and a Ctrl-F-findable quotation of no more than 15 words to each factual claim, with a table for verified and unverified items. It is useful for traceable answers but cannot replace web access; paywalls, JavaScript pages, and changing links can still limit checking.

Critic Gauntlet runs parallel adversarial critiques from model families such as Claude, Codex, Grok, and Gemini, then synthesizes major gaps, steel-manned alternatives, assumptions, six-month consequences, and a recommendation. It is aimed at expensive-to-reverse architecture choices, not routine bug fixes or refactors. Extra critics require their CLIs, APIs, and credentials.

This two-skill bundle can refresh Volcengine official price CSVs through browser automation and generate a cloud-migration quote workbook from Alibaba Cloud, Tencent Cloud, AWS inventories, or customer needs. It maps more than 80 product types, offers three sheet strategies, and includes validators. Full price refresh needs a browser; procurement APIs, discounts, and live competitor pricing are out of scope.

ReelForge is a local video and caption studio that extracts Whisper word timestamps, creates synchronized TikTok-style captions and translations, and adds motion graphics, charts, maps, social cards, and b-roll. It can also use TTS, background removal, and website-to-video workflows, with HyperFrames, whisper.cpp, and FFmpeg under the hood. Some stock-media or HeyGen features may need APIs.

PPT Design is a Claude skill for editable 16:9 PPTX decks built from a design system. It supports proposals, reports, lectures, and guides, with 45 layouts plus native tables and charts, and pays attention to Korean slide readability. It requires python-pptx and suitable fonts; fonts are not embedded, and rendered checking needs PowerPoint or LibreOffice.

Suno Skill turns a song idea into paste-ready Suno text: a vocals-first style box, tagged lyrics, settings, and notes in Dutch or English. Its five-step flow covers concept, style, lyrics, meter-preserving humanization, and a quality checklist, with pronunciation fixes, exclude tags, sliders, personas, and extension controls. It is text-only and uses no API or browser automation.

AnySurfer Audit is an unofficial Claude skill for evaluating HTML, CSS, JavaScript, Figma, Penpot, screenshots, wireframes, or URLs against WCAG 2.2 A/AA and the AnySurfer approach. Reports organize findings by accessibility principle, severity, contrast, estimated label level, and fix order. Incomplete inputs limit visual and runtime conclusions.

The Council is a Claude Code panel for validating one idea or comparing two options. Permanent critics and cast domain specialists research in parallel, while a Chair produces a structured verdict with objections, blocking conditions, and next actions. Runs offer Quick, Standard, and Deep effort levels, resumable sessions, a logbook, and Markdown or HTML reports; extra critics need their own CLIs or APIs.