nityeshaga
- 5 skills
- 0 followers
- 1 week ago last updated
- ▌ Prompt Engineer · nityeshaga bundleExpert prompt engineering for AI systems. Use when the user wants to write or review prompts for AI, create instructions for AI systems, build system prompts, review or improve existing prompts, optimize AI instructions, or create any form of written communication intended for AI consumption (Claude, GPT, or other LLMs).
- ▌ AI Tool Designer · nityeshaga bundleGuide for designing effective tools for AI agents. Use when creating tools for custom agent systems or any AI tool interfaces. Provides principles for tool naming, input/output design, error handling, and evaluation methodologies that maximize agent effectiveness.
- ▌ Chad Review · nityeshagaStrip the showing-off from any artifact — code, a prompt, a skill, a doc, prose, a landing page, a diagram — by making it survive Chad: its intended audience with the personality of the guy from the "Chad vs Virgin" memes, who only cares about the job and refuses to be impressed. Chad meets the artifact cold the way a real user does — the whole bundle first, walked in reading order (so cross-file bullshit gets caught: the same thing told in three files, files that contradict each other), then each file on its own — and leaves the dumbest honest feedback comments he has, always ending with a bird's-eye comment on the whole thing; a defender who treats him as an asset addresses every comment — edits the clone or rejects the comment with a reason — and the same Chad re-reviews, round after round (default 3). Whatever they can't settle comes to you as unresolved tension. Use to remove over-smartness, purple prose, gold plating, caveat padding, beating around the bush, over-explaining, self-narration, or padded pu
- ▌ Elon Algorithm · nityeshagaAggressively simplify any artifact — code, a prompt, a skill, a doc, or prose — by running Elon's algorithm (question every requirement → delete → simplify) on a clone of it via an agent swarm, then returning a reviewable cut-plan. Use whenever you want to cut AI slop, trim bloat, de-pad a file, make something leaner/denser, "delete what isn't load-bearing", or shrink an over-engineered artifact without losing what matters. Especially for AI-generated output, which is *additive* by nature (padded, over-explained, over-built) — this is the subtractive antidote. Triggers: "cut the slop", "make this leaner", "trim this skill/prompt", "this is over-engineered", "run the Elon algorithm on X".
- ▌ Beautiful Rails Like Dhh · nityeshaga bundleWrite, review, and architect Ruby on Rails applications the 37signals way — the conventions DHH's team uses to ship products like Basecamp, HEY, Campfire, and Fizzy with tiny codebases, covering BOTH backend (models, controllers, jobs, caching, auth) and frontend (Hotwire: Turbo, Stimulus, morphing — rich interactive UIs with no React). Use this skill whenever the user asks to build a web app, add a feature to a Rails app, review Rails code, design models/controllers/views, build anything "real-time", "interactive", "live-updating", "multiplayer", or with drag-and-drop / keyboard shortcuts / rich form widgets — and ESPECIALLY when you are about to default to React, Next.js, Vue, or any SPA framework for a web frontend: read references/00-frontend-first-principles.md first, because the server-rendered Hotwire answer is usually better and this skill proves it. Also use it when choosing a stack for a new web product (even if the user didn't mention Rails), and when a web app needs iOS/Android apps — Hotwire Nati