Guardado en:
| Autor principal: | |
|---|---|
| Formato: | Preprint |
| Publicado: |
2026
|
| Materias: | |
| Acceso en línea: | https://arxiv.org/abs/2604.02544 |
| Etiquetas: |
Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
|
| _version_ | 1866911564350619648 |
|---|---|
| author | Borysenko, Oleksii |
| author_facet | Borysenko, Oleksii |
| contents | The rapid adoption of AI coding agents and AI assistant web services is fundamentally changing how developers discover, consume, and interact with technical documentation. This paper studies that transformation across three interconnected dimensions: documentation accessibility, content analytics, and feedback systems. We present an empirical study of HTTP request fingerprints from nine AI coding agents (Aider, Antigravity, Claude Code, Cline, Cursor, Junie, OpenCode, VS Code, and Windsurf) and six AI assistant services (ChatGPT, Claude, Google Gemini, Google NotebookLM, MistralAI, and Perplexity) accessing a live developer documentation endpoint, revealing identifiable behavioral signatures in HTTP runtime environments, pre-fetch strategies, User-Agent strings, and header patterns. Our study shows that AI agent access compresses multi-page navigation into a single or two requests, making traditional engagement metrics - session depth, time-on-page, click path, and bounce rate - unreliable indicators of actual documentation consumption. We discuss practical adaptations for developer portal teams, including tokenomics-aware documentation design, adoption of emerging machine-readable standards (AGENTS.md, llms.txt, skill.md, agent-permissions.json), MCP server-based feedback channels, and analytics instrumentation for AI referral traffic. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2604_02544 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | Developer Experience with AI Coding Agents: HTTP Behavioral Signatures in Documentation Portals Borysenko, Oleksii Software Engineering The rapid adoption of AI coding agents and AI assistant web services is fundamentally changing how developers discover, consume, and interact with technical documentation. This paper studies that transformation across three interconnected dimensions: documentation accessibility, content analytics, and feedback systems. We present an empirical study of HTTP request fingerprints from nine AI coding agents (Aider, Antigravity, Claude Code, Cline, Cursor, Junie, OpenCode, VS Code, and Windsurf) and six AI assistant services (ChatGPT, Claude, Google Gemini, Google NotebookLM, MistralAI, and Perplexity) accessing a live developer documentation endpoint, revealing identifiable behavioral signatures in HTTP runtime environments, pre-fetch strategies, User-Agent strings, and header patterns. Our study shows that AI agent access compresses multi-page navigation into a single or two requests, making traditional engagement metrics - session depth, time-on-page, click path, and bounce rate - unreliable indicators of actual documentation consumption. We discuss practical adaptations for developer portal teams, including tokenomics-aware documentation design, adoption of emerging machine-readable standards (AGENTS.md, llms.txt, skill.md, agent-permissions.json), MCP server-based feedback channels, and analytics instrumentation for AI referral traffic. |
| title | Developer Experience with AI Coding Agents: HTTP Behavioral Signatures in Documentation Portals |
| topic | Software Engineering |
| url | https://arxiv.org/abs/2604.02544 |