Ilm-Alan

Build a frontend with a deliberate visual direction held through palette, typography, structure, and texture, with on-screen strings that name real information rather than fabricated demo data or filler labels. Use when building or restyling a frontend.

Rewrite prose to remove signs of LLM authorship and inject specific, varied, opinionated human voice. Use when humanizing AI-drafted text. Skip for code, commits, configs, and prose you are currently drafting (apply inline instead).

Create a project spec for AI agents to work from autonomously.

Read, search, and send Discord messages in servers where the configured bot is installed. Use when asked to inspect a Discord channel, search server history, post a message, or interact with a DM addressed to the bot.

Implement a UI screen, page, or component to match a reference image — mockup, screenshot, or photographic render — with visual fidelity that survives the target's aspect ratio, while keeping every rendered value and every affordance honest: real data and working features only, never mockup values transcribed into code, placeholders fabricated to fill the layout, or stubbed controls that look real but do nothing. TRIGGER when: user hands you a design image, mockup, screenshot, photographic reference, Figma frame, or exported comp and asks to build, implement, match, recreate, or "make it look like" it. SKIP: originating a visual direction with no reference image (use the frontend-design skill); pure copy or content edits; backend-only work with no screen to match.