{"id":"ryanmac-agent-twitter-client-mcp","name":"agent-twitter-client-mcp","af_score":67.8,"security_score":52.2,"reliability_score":41.2,"what_it_does":"Provides an MCP (Model Context Protocol) server that exposes Twitter/X actions (read tweets, search, post/like/retweet/quote, follow/profile data) to AI agents, including optional Grok chat/realtime features via the underlying agent-twitter-client integration.","best_when":"You want agent-to-Twitter functionality via MCP tools (e.g., Claude Desktop) and can supply the necessary Twitter credentials/cookies, accepting that Twitter/Grok behavior may vary and requires ongoing maintenance.","avoid_when":"You need strict OAuth scopes-based least privilege, domain-verified enterprise security controls, or you cannot securely manage Twitter credentials/cookies in an environment where agents may access them.","last_evaluated":"2026-03-30T15:29:55.918291+00:00","has_mcp":true,"has_api":false,"auth_methods":["cookie-based authentication","username/password authentication","Twitter API v2 credentials (API key/secret and access tokens)"],"has_free_tier":false,"known_gotchas":["Cookie authentication requires cookies in a specific JSON array format with proper domains; expired cookies will break auth.","Twitter/Grok may trigger Cloudflare/captcha especially for username/password auth (docs warn of this).","Grok functionality requires agent-twitter-client v0.0.19+; the package description/README notes current basic functionality uses v0.0.18 and demo flags may install v0.0.19.","Grok has rate limits (non-premium: 25 messages/2 hours per README) and the MCP returns rate limit info only when limits are reached.","Agent workflows may repeatedly trigger mutation tools unless the agent host implements throttling/guardrails."],"error_quality":null}