agent-ecosystem

Agent Traffic Classifier — Data skill for Claude Code

Data community

Classify web traffic from server log entries and header capture into human, AI agent, AI crawler, AI assistant, programmatic, and other bot categories.

How to install Agent Traffic Classifier

This entry records only its repository, not the path inside it, so there is no exact command to give. Open agent-ecosystem/agent-traffic-classifier and copy the folder into ~/.claude/skills/, or the file into ~/.claude/agents/.

What Agent Traffic Classifier does

Classify web traffic from server log entries and header capture into human, AI agent, AI crawler, AI assistant, programmatic, and other bot categories. No analytics service or middleware required.

Alternatives in Data

  • HTML Video — Programmatic video for coding agents — HTML to video on your laptop 4.5k ★
  • Memex — Open-source, local-first AI journal app for iOS and Android 697 ★
  • Local News Site — Publish a local news article as a standalone Elementor page, with Google News-grade structured data and image 648 ★

README

CI npm

agent-traffic-classifier

Classify web server traffic into humans, bots, AI agents, and programmatic clients. Built for analyzing access logs and HTTP header signals to understand how AI agents interact with your site.

**Status: Early development (0.x)** This library is currently in early iteration and does not promise stability guarantees.

What it does

Given access log entries (and optionally HTTP header signals), the library:

  1. Classifies each request by user-agent into a category: human, AI crawler, AI assistant, AI search, coding agent, search crawler, SEO bot, monitoring, social preview, feed reader, programmatic client, or unknown
  2. Detects AI agents that use standard browser user-agents (Claude Code, Cursor, Kiro, Gemini CLI) via HTTP header heuristics and signal attribution
  3. Identifies agent frameworks by Accept header patterns, missing browser security headers, and conversation-tracking headers, even when the specific agent is unknown
  4. Attributes country-level intelligence to unidentified agents using IP ranges from Regional Internet Registries, enabling suspected identification (e.g., "Kimi / Doubao / DeepSeek (suspected)" for Chinese AI assistants)
  5. Clusters sessions by correlating signal data with access logs to reclassify traffic that would otherwise look human
  6. Cross-references programmatic traffic with signal data to upgrade HTTP client requests (httpx, undici, etc.) to agent when they share IPs with known agent signals
  7. Detects proxy-based agents (like Cursor) via a duplicate-request heuristic: same path + UA from di