X MCP & Agent Tooling

Appearance

skill.md Fresh ​

The skill.md file follows the agentskills.io specification and describes what AI agents can do with the X API. While llms.txt is a directory of pages, skill.md is a capability summary - it lists specific actions, required inputs, and constraints so agents can use the API more reliably.

What's included ​

  • Capabilities - what agents can accomplish (search posts, create posts, manage users, etc.)
  • Skills - specific actions organized by category with required parameters
  • Workflows - step-by-step procedures for common tasks
  • Context - background on authentication, rate limits, and architecture
bash
# Fetch the skill file
curl https://docs.x.com/skill.md

llms.txt vs skill.md ​

flowchart LR
    subgraph llms.txt
        A[A directory of pages]
        A --> A1[Titles]
        A --> A2[URLs]
        A --> A3[Descriptions]
    end
    subgraph skill.md
        B[A capability summary]
        B --> B1[Actions]
        B --> B2[Required inputs]
        B --> B3[Constraints]
    end

llms.txt answers "what pages exist?". skill.md answers "what can I do, and what do I need to do it?". For action-oriented agents, skill.md produces more reliable tool use because it is framed around capabilities rather than documents.

Discovery endpoints ​

Agents can discover skill files programmatically via the well-known endpoints:

bash
# Discovery endpoint (agent-skills 0.2.0 spec)
curl https://docs.x.com/.well-known/agent-skills/index.json

# Original discovery format
curl https://docs.x.com/.well-known/skills/index.json

Add to any skills-capable agent ​

You can add X API capabilities to any agent that supports the skills CLI:

bash
npx skills add https://docs.x.com

See also ​