Shape agent capabilities into your workflow.

video-bridge retrieves Bilibili subtitles when available and falls back to audio transcription, then turns short caption fragments into timestamped sections, full transcripts, and notes. SQLite FTS5 enables search across videos, with Obsidian-friendly Markdown as an option; subtitle access needs a logged-in session and fallback transcription depends on ffmpeg, yt-dlp, and podcast-bridge.

Seaworthy performs a static, no-network scan before deployment for common security mistakes in AI-built web apps, including exposed secrets, missing row-level security, unauthorized mutations, risky CORS, debug mode, open admin routes, and writable storage. It returns a ship-or-do-not-deploy decision with file-level findings; it is not a penetration test and primarily targets Next.js and Supabase scenarios.

Give Ultimate Idea Watcher a sentence, rough notes, screenshots, or an existing codebase, and it preserves the original idea while building a structured product brief around it. The skill researches the problem, users, and alternatives, defines the product architecture, then splits work into dependency-aware phases with ready-to-use designer and engineering prompts. It also maintains traceability and gap checks across security, privacy, accessibility, and later plan changes.

Start with confirmed bookings, scattered notes, or no reservations at all. The workflow compares route and pacing options, confirms the itinerary structure, then verifies time-sensitive details such as opening hours, fares, events, and transport before assembling a mobile-first single-file HTML handbook and printable version. Unbooked items remain clearly labeled, with a prioritized list showing what to reserve first.

Install repaint with `npx skills` and ask a compatible coding agent to build or redesign an interface. Its workflow researches real references, commits a concrete token system before markup, then renders the result and iterates through visual-slop, interaction, accessibility, and public-page SEO audits. The package is self-contained, with a curated reference library and 23 evaluations, and does not depend on other skills or agents.

handover-navigator reads documents, emails, notes, spreadsheets, and meeting records to create a handover summary, open-task tracker, first-30-day onboarding plan, and questions for missing context. It links each claim to a source, turns unknowns into questions instead of guesses, and normalizes inconsistent task states. Original material stays read-only while results go to a separate directory.

meeting-minutes-skill sends meeting audio to Alibaba Cloud OSS and Tongyi Tingwu for transcription and speaker separation, optionally extracts text from presentation files, and produces formal Chinese minutes with action items and review notes before exporting a Word document. It keeps the transcript as the factual source and warns against inventing names, budgets, or promises. OSS and Tingwu credentials are required.

claude-vastai-skill provides a Claude Code and terminal control plane for searching, renting, watching, using, syncing, and destroying Vast.ai GPU instances. Each rental can carry an auto-destroy deadline guarded by a watchdog, with cost and orphan checks; the CLI wraps the official Vast.ai tool and still requires an API key.

humanizer-fr rewrites pasted French text or a file to remove formulaic language-model habits, translated phrasing, and French typography mistakes. It adapts 33 patterns from the English humanizer and can calibrate the voice against a sample of the user’s own writing. The skill is prompt-only: it edits text but contains no executable code, making it suitable for Claude Code or OpenCode.

github-search-skill searches GitHub across repository, code, REST, GraphQL, and CLI angles to uncover useful projects that star rankings may hide. It includes a dead-end guide, code-search examples, and a batch shell helper with deduplication and rate-limit handling. The workflow uses GitHub’s official API rather than scraping and requires an authenticated gh CLI plus jq.

mask rewrites bland copy in the voice of a chosen character by drawing on linguistic fingerprints such as self-reference, catchphrases, sentence rhythm, and diction. You can use it for a pasted line or draft and switch among example voices such as Sherlock Holmes, Yoda, or Shakespeare. The goal is a consistent, recognizable character voice rather than a shallow word swap.

questionnaire-analysis-skill guides bilingual survey and scale research from questionnaire design and data cleaning through reliability and validity, group tests, regression, SEM, mediation, moderation, and APA reporting. It packages methodological references, result-sentence templates, checklists, and an offline threshold reference. The approach is tool-neutral but centers on SPSS and AMOS, with R examples for key steps.

claude-loop converts “make it better” into a project-specific scorecard and a repeatable cycle: score real signals, triage the weakest link, fix one slice, verify independently, and score again. Its init, score, run, and status modes keep the rubric and checkpoint state in .loop/. The loop stops after the target score is reached twice, with 90 as the default, instead of running indefinitely.

forex-sentiment-monitor combines GDELT and RSS with optional news APIs to score sentiment for foreign exchange, precious metals, and crypto symbols. VADER supplies a fallback score, while Claude Code can add industry-context review, causal transmission chains, and divergence explanations. Reports are written as Markdown or HTML with source attribution; the default free sources need no API key, and the output is for research rather than investment advice.

paper-ingest resolves a paper URL, title, or DOI into metadata, extracts full text from the PDF, classifies the work, and renders a detailed Obsidian note. It can push references into Zotero, detect duplicates, and collect PDFs for later attachment through WebDAV. The skill expects a local config containing service credentials, so it fits researchers who already use Zotero and Obsidian.

Describe an image or provide a weak existing prompt, and this Claude skill will ask up to three questions only when a critical choice is missing. It then builds one prompt through a ten-part structure covering composition, subject, lighting, palette, camera or typography, finish, and references, with adapters for GPT Image 2 and Nano Banana. It writes prompt text only and does not call an image-generation API.

subtitle-to-study-page converts SRT, VTT, or TXT subtitles plus a video link into a self-contained interactive HTML study page. It organizes chapters with takeaways, analogies, study questions, detailed summaries, cleaned translations, and a jump-to-source link, then adds a searchable timeline and chapter-aware toolbox. The generated page turns long-form video subtitles into structured chapter notes for review.

Point the skill at a React or Next.js codebase to build a guided, theme-matched preview that visitors can click through before creating an account. It reconstructs only the required screens on in-memory mock data, adds a preview route and launch button, and audits the generated code to keep API clients, authentication, secrets, and real backend logic out of the demo.