writing▌

Expert-level guidance for writing publication-ready papers targeting NeurIPS, ICML, ICLR, ACL, AAAI, and COLM. This skill combines writing philosophy from top researchers (Nanda, Farquhar, Karpathy, Lipton, Steinhardt) with practical tools: LaTeX templates, citation verification APIs, and conference checklists.

This skill provides comprehensive guidance for writing excellent academic and research papers across all disciplines. It covers structure, style, argumentation, and best practices from initial planning through final revision.

Craft distinctive long-form content that matches a specific voice or brand style, not generic AI output. \n \n Captures and applies existing voice patterns from articles, newsletters, or style guides to maintain consistency across blog posts, guides, tutorials, and launch announcements \n Enforces concrete-first writing: leads with examples, screenshots, code, or data before explanation, avoiding generic openings and filler transitions \n Bans hype language and unsourced claims; requires factual

Write with clarity and force. This skill covers what to do (Strunk) and what not to do (AI patterns).

Edit and improve articles by restructuring sections, enhancing clarity, and tightening prose.

CSF3-compliant Storybook stories that showcase component variations with type safety and consistent structure. \n \n Enforces Component Story Format 3 with TypeScript type annotations, default meta exports, and story object syntax \n Covers story organization patterns for buttons, forms, layouts, data-driven components, and responsive designs \n Includes best practices for extending stories, using decorators for context, and adding descriptive parameters for documentation \n Identifies anti-patt

Apply Strunk's timeless writing rules to improve clarity and professionalism across all prose. \n \n Covers 18 core rules spanning grammar, punctuation, composition principles, and common usage mistakes \n Emphasizes active voice, positive framing, concrete language, and ruthless word elimination as key clarity drivers \n Designed for documentation, commit messages, error messages, UI copy, reports, and any prose humans will read \n Includes token-efficient subagent dispatch pattern: write a dra

Framework for writing job descriptions that attract the right candidates by defining business impact. \n \n Start with 12-month success: define what changes in the business after the hire, not a list of daily tasks \n Identify the spike: determine one or two areas where the candidate must excel, rather than seeking a generalist \n Replace proxies with outcomes: frame requirements around specific capabilities and progress, not years of experience or tool lists \n Use polarizing language intention

Techniques for writing effective fuzzing harnesses across languages and frameworks. \n \n Covers harness fundamentals for C/C++, Rust, Go, and Python, including entry point signatures, input validation, and structured data extraction using FuzzedDataProvider and arbitrary crate \n Includes step-by-step guidance from identifying entry points through testing and iteration, with practical patterns for integer casting, complex inputs, and interleaved fuzzing of multiple operations \n Provides tool-s

Define patterns and messages that trigger Claude warnings or blocks on specific agent actions. \n \n Supports four event types (bash commands, file edits, agent stop, user prompts) plus an \"all\" catch-all, each with regex or multi-condition matching \n Rules use markdown files with YAML frontmatter stored in .claude/hookify.{name}.local.md ; changes take effect immediately on next tool use \n Two matching modes: simple pattern field for single-condition rules, or advanced conditions array for