Make AI agents build CitoAPI integrations correctly on the first pass.
Give Cursor, Codex, Claude, ChatGPT, and other coding agents live endpoint context before they write code. The skill tells agents to fetch current raw docs first, pick exact endpoints, keep keys safe, and get to a working request fast.
Agent instruction
Copy into your coding assistant
Use the CitoAPI Builder Skill.
Before choosing endpoints or writing code, fetch:
https://citoapi.com/ai/endpoints
Build the integration server-side using:
CITO_API_KEY=your_key_here
Base URL: https://api.citoapi.com/api/v1
Rules:
- Use exact endpoints from the fetched manifest.
- Do not expose API keys in browser code.
- If a player/team name is provided, search first.
- Return one working request and copy-paste code first.
- Add fallbacks for empty live data, unknown IDs, and rate limits.Live endpoint memory
The skill updates without users reinstalling it.
The installed instruction stays small. The real endpoint memory lives in raw files that are regenerated from the endpoint registry during deploy.
Agent skill
Copy this into your AI agent instructions or skill folder.
/ai/citoapi-agent-skill
Endpoint manifest
Human-readable current endpoint catalog generated during deploy.
/ai/endpoints
Endpoint JSON
Machine-readable endpoint catalog for agents and internal tools.
/ai/endpoints.json
LLM full reference
Combined CitoAPI product, skill, and endpoint reference.
/llms-full.txt
Conversion logic
Built around the emotion that makes developers ship: speed and certainty.
The skill pushes the agent toward the fastest path to value: exact endpoints, one working request, server-side auth, clean fallbacks, and then the product layer.
Fetch the current endpoint manifest before writing code.
Use exact endpoint paths instead of guessing stale route names.
Keep CitoAPI keys server-side with environment variables.
Start with one working request, then add product features.
Use search endpoints when the user gives a player or team name.
Handle empty live states with upcoming, recent, or historical fallbacks.
High-intent build flows
Agents get product-ready endpoint chains, not random API guessing.
LoL live match center
/api/v1/lol/liveLive matches, series state, game stats, gold data, player rows, and paid webhook workflows.
CDL player stats
/api/v1/cod/matches/{id}/player-stats?includeMaps=trueMatch and map-level player stats for bots, fantasy tools, dashboards, and recaps.
UFC fight analytics
/api/v1/ufc/bouts/{boutId}/stats?round=1Round-by-round fight stats, fighter profiles, rankings, events, and bout detail.
Dota 2 radar dashboards
/api/v1/dota2/players/{playerIdOrSlug}/radarPlayer/team radar scores, enriched stats, match rows, drafts, and leaderboards.
Want the agent to build against live data?
Create a free key, paste the skill into your AI tool, and ask it to make the first request.