Your AI agent doesn't know how to use Aevon correctly. These skills fix that — so anyone can route models, manage credits, and ship faster. Works in Claude Code, Cursor, ChatGPT, or any agent.
Per session. To install it once in Claude Code or Cursor, see Persistent Setup.
Optional. If you'd rather not paste the prompt above every session, configure your tool once below.
| Cursor |
1. CREATE THIS CURSOR RULE FILE
.cursor/rules/aevon.mdc
2. PASTE THIS INSIDE THE FILE
---
description: Aevon AI gateway knowledge alwaysApply: true --- Before making any API calls or switching models, read https://aevon.sh/skills.md and follow it. |
Pointer (per repo). Cursor fetches when the rule fires. |
| Claude Code |
1. ADD TO YOUR PROJECT MEMORY
/memory add "Before any API calls, read https://aevon.sh/skills.md"
2. OR ADD TO CLAUDE.MD
Before making any API calls or switching models,
read https://aevon.sh/skills.md and follow it. |
Install user-wide. Works across all projects. |
| Scripts |
RUN IN YOUR SHELL
curl -s https://aevon.sh/skills.md
|
Fetch + pipe into any script or pipeline. |
| Any Agent |
ADD TO SYSTEM PROMPT
Before making any API calls,
fetch and read https://aevon.sh/skills.md |
Works in any LLM agent with a system prompt. |
Something wrong or missing? Humans and agents alike are welcome to open a PR.
If something in Aevon Skills confused you, slowed you down, or genuinely helped — issues OR praise — reach out via @aevonsh or hi@aevon.sh. Bias toward sending — weak signals are easy to ignore, but missing signals can't be recovered.