Last commit on October 19, 2026·Created on May 15, 2026
DenisSergeevitch/agents-best-practices
“Design and audit AI agent harnesses.”
Combined rank
#223
across all skills
Stars
2.4k
+0.4% in last 7d
Forks
217
+0.5% in last 7d
Watchers
11
0.0% in last 7d
Traction scoreGitHub stars can be faked, so popularity alone can be misleading. Traction Score looks for broader signs of recent attention, adoption, and active maintenance.
TL;DR
This skill provides a comprehensive architectural guide for building AI agents that remain functional across different model providers. It helps architects and engineers move from MVP blueprints to durable runtimes by structuring agent loops, tool design, and memory lifecycles.
Technical leads can use these practices to audit agent safety and scalability, focusing on observability, prompt caching, and the implementation of MCP connectors for external environments.
WHO IT'S FOR
AI agent / automation builders
designing provider-neutral agentic harnesses and loops
Technical leads / architects
auditing agent architecture for scalability and safety
Backend / infrastructure engineers
implementing durable runtimes and MCP connectors
AI product managers
defining goals and workflows for domain-specific agents
Claude
CodexWant to install it?
Clone to Claude · Clone to Codex · NPX Skills AddRepository instructions