Dev scans 5,000 MCP servers daily: 10% dead, 30% auth-gated, then ClaudeBot shows up
art2meta · reddit · 2026-10-27
An indie developer built lowdown, a read-only daily probe of 5,000 remote MCP servers from the official registry (initialize + tools/list only, no credentials), with open-source code and CC BY 4.0 data.
Early findings:
- 10% of servers unresponsive for three consecutive days
- 30% require authentication
- Tool definitions change surprisingly often; some so frequently that hash diffs alone are meaningless
Unexpected twist: ClaudeBot made 6,191 requests against the API across 5,262 targets in seven days, exposing a weakness (30-day aggregation overloading the DB), now fixed with precomputed views and explicit errors. The thesis: agents shouldn't pick tools based on names/descriptions alone — availability, auth requirements and definition churn are the evidence that matters.
More from coding & agent
- Grok bot makes its first real purchase, spots hidden prompt injection at checkout and ignores it — msg · 2026-10-27
- Coding agents now make 60% of Vercel deployments, up from 3% in January — SuB8u · 2026-10-27
- Delta, Chase and Amazon are blocking AI agents — and the blockers may lose, predicts 6-stage war — alvelda · 2026-10-27
- Instinct launches Link balance to let agents spend within set budgets — jeff_weinstein · 2026-10-27
- Production lesson: a fluent AI translation can still be a broken document — WirelessLife · 2026-10-27
- One-prompt test: Grok Bot nails a premium DTC supplement landing hero — azed_ai · 2026-10-27