Chaining Grok Bot and Codex to detect and mass-block AI reply accounts
vista8 · x · 2026-10-27
A hands-on workflow chaining two agents to auto-detect and block AI reply spam: Grok Bot scans 250 replies from the past week and flags likely AI accounts (about 15–20%), scoring cross-post template reuse, new low-follower accounts and stylistic patterns, then writes a structured report to a Feishu doc; Codex then takes the doc and drives Browser Use to visit and block each account. No code required, and the author offers an appeal channel for false positives.
More from coding & agent
- 'Proactive' AI agents are just cron jobs poking your agent every 5 minutes — cspenn · 2026-10-27
- Thoughtworks: don't ask AI to write user stories yet — ask it to find the decisions nobody made — rseroter · 2026-10-27
- Ex-GitHub CEO Dohmke Launches Entire, a Git Forge Built for Agent-Human Collaboration — hboelman · 2026-10-27
- Tool calling is still just a prompt: gateway lessons from Qwen3's chat template — Main-Fisherman-2075 · 2026-10-27
- Supabase adds scoped Personal Access Tokens for least-privilege CI and coding agents — dshukertjr · 2026-10-27
- Licensed law firm Crosby launches Slack paralegal agents in beta — graceisford · 2026-10-27