Loading...
Loading...
Found 22 Skills
arxiv 논문을 체계적으로 분석하여 Obsidian 서베이 노트를 생성한다. arxiv 페이지/ID를 입력받아 (1) Paper Digest 스타일 핵심 요약, (2) Iterative Writing 원칙 기반 세부 정리, (3) 플로우 다이어그램 포함 Obsidian 페이지를 생성한다. TRIGGERS: "논문 서베이", "paper survey", "arxiv 정리", "논문 노트", "서베이 페이지", "survey note"
Writes or rewrites README.md files tailored to the project type (CLI, library, app, framework, monorepo, or skill bundle). Discovers project context, selects the right structure, writes section by section, and validates against quality checks. Use when creating a README, writing a README from scratch, rewriting a bad README, bootstrapping project documentation, or asking "write a README for this project."
Use when generating 50+ structured items with parallel Claude Code subagents and merging outputs into one file.
Use when creating, modifying, or refactoring Claude Code skills that require structured multi-agent review and quality validation
Design, audit, and improve analytics tracking systems that produce reliable, decision-ready data. Use when the user wants to set up, fix, or evaluate analytics tracking (GA4, GTM, product analytics, events, conversions, UTMs). This skill focuses on measurement strategy, signal quality, and validation— not just firing events.
Invoke this skill when a user shares test code and questions whether it actually works as intended — not to run or fix the test, but to evaluate whether the test has real value. Triggers on: "is this test any good?", "would this catch a real bug?", "this test always passes — is that normal?", "review these tests before I commit", or "does this test verify anything meaningful?". Also triggers when someone suspects a test is useless, wants a pre-commit quality gate, or is unsure if an auto-generated test is worth keeping. The core question this skill answers: "Would this test fail if the feature broke?" If not, the test gets rejected. Do NOT use for generating new tests, fixing failing tests, or exploring application features.
Use this skill to migrate course catalog data from external sources (CSV, PDF, website) and bulk-create Learning and LearningCourse records in Education Cloud. Triggers when a user wants to import courses, parse course catalog documents, or set up a course catalog for an academic term. Handles multi-source data ingestion, intelligent interpretation of course structures (lecture+lab linked courses via IsLinkedOnly flag, credit hours, course types), batch creation via sObject REST API, data quality validation, conflict resolution for existing courses, and gap identification. Explains what's missing (descriptions, types) and requests additional sources. Pauses for clarification on multi-component courses and conflicting data. DO NOT TRIGGER for creating individual Learning Course records manually, for creating Learning Programs (different workflow), or for enrollment/registration workflows.
Use when validating code quality after modifications. Runs sniper agent in isolated forked context for clean, fast validation.
Generate storyboards for AI short dramas with a three-layer structure: Segment (one video generation, ≤15 seconds) → Shot (2–5 second cuts within a segment, assigned to script beats) → Storyboard Frame (one keyframe per cut: main storyboard frame pinned at 0.00s, sub-frames pinned at their respective cut points). Each segment comes with a MiniMax H3 video prompt (official default is English, line-broken per shot, promptLang can be switched to Chinese): alignment instructions and [Shot k] cut timings are derived from the storyboard structure and verified character by character; dialogue is placed character by character into <d> blocks (writing specifications are internalized in references/h3-prompt.md, no external skill dependencies). Outputs include storyboard.json + Markdown + single-page review report (storyboard rhythm bar / episode-by-episode storyboard table / generation batch sheet / dubbing alignment sheet, with exportable JSON). No images are generated — the deliverables are visual prompts for each frame and the reference images that should be attached to them. All 18 quality gates are deterministically checked by scripts (the 18th shot-recipe gate is optional: it is only checked when the shot-recipes card library is mounted, otherwise it is explicitly skipped). One-click export of H3 production packages via export (prompts per segment + storyboard frame list in Picture order). Zero dependencies, zero API keys, uses current session quota. Use when requested to create storyboards, generate storyboards, produce shot lists, edit cuts, or work on storyboards for AI short dramas.
Builds clean brand-colour logos for every company, app or tool named in a video. Lists the names from the edit's transcript, fetches each brand's own logo (SVG first, else a large PNG) with the editor's logo code, trims it, makes a transparent 512 px PNG beside the SVG, checks each one (official, transparent, sharp, brand colours) and fixes the failures, then notes the logos named together so the edit shows them side by side. Writes the files style-edit already reads (images/logo-<brand>.svg|png) plus logos.json. Use when the user says "make a logo pack", "get the logos for this video", "official logos for the apps I mention", "the logos look fake" or "logos in their real colours", or before style-edit on a video that names two or more products. Not for moving or swapping one card in a rendered edit (overlays), and not for screenshots of a page (style-edit captures).