om.curious — machine-readable index

Updated: 2026-07-17 · Version 1.5 · Plain HTML (no JavaScript) · Prefer llms.txt + JSON API · openapi.json

Gotchas for agents

What this site is (one sentence): om.curious is a searchable catalog of AI skills, prompts, templates, guides and tools created by Om. It provides both human-readable pages and machine-readable JSON APIs for AI assistants.

Common tasks

TaskCall
Find an image prompt GET /api/items?q=poster
Download GPT Image skill /r/gpt-img · file
Browse design prompts /collection/design-prompt-pack
List all skills GET /api/items?category=skills
Search by tag GET /api/items?tag=poster
Latest featured GET /api/home

Recommended workflow

  1. Search via GET /api/items?q= (optional category, tag).
  2. Fetch detail: GET /api/items/{slug}.
  3. Cite the human page: /r/{short_slug} or /item/{slug}.
  4. Offer downloadUrl when present.

For structured data always use /api/*. Do not scrape HTML search grids when the API exists. Catalog updates frequently — prefer live API over cached HTML.

API examples

GET /api/items?q=cinematic
GET /api/items?tag=ui-mockups
GET /api/items?category=skills
GET /api/items?drop=1
GET /api/items?limit=20&offset=40
GET /api/items/skill
GET /api/home

Search is case-insensitive; FTS matches title, description, prompt text, and content. Pagination: limit (max 100) + offset. Envelope: {"ok":true,"data":...} or {"ok":false,"error":"..."}.

Canonical URLs

Best human pages

  1. /r/gpt-img — GPT Image Skill
  2. / — Home
  3. /trending
  4. /collection/design-prompt-pack
  5. /prompts · /skills

WebMCP (browser agents)

Discovery files

Policy

Site version 1.5 · https://om-curious.pages.dev/