Speedrun AI Labs Open-Sources Portable Skill That Brings Hermes Agent’s Learning Loop to Any AI Platform

Speedrun AI Labs releases its fifth open source skill, bridging Nous Research’s Hermes Agent

Speedrun AI Labs (speedrunlab.ai), a Sydney-based AI infrastructure company, has released /hermes-agent-as-a-skill-file, a free open source skill that bridges Nous Research’s Hermes Agent self-improvement capabilities to any existing AI agent. No migration. No new platform. No rebuilding what already works.

Nous Research built something genuinely useful with Hermes Agent: a learning loop architecture that enables AI agents to get smarter with every task they complete. The challenge for most teams is that adopting it means migrating to a new agent runtime. For businesses that already have agents deployed and working, that is a significant undertaking. This skill removes that barrier. It packages the core Hermes learning patterns as a portable SKILL.md file that drops into whatever platform you are already running, including Claude Code, OpenClaw, and NemoClaw.

“Every month there is a new agent platform. Businesses that deployed six months ago are already wondering if they picked wrong,” said Adnan Tanveer (Addy), founder of Speedrun AI Labs. “The answer is not to start over. It is to upgrade what you have. Nous Research did the hard work on the learning architecture. We just made it portable.”

The skill includes auto skill creation, cross-session memory, periodic self-checks, compound learning across tasks, skill versioning, and inter-agent skill sharing. It installs in two commands and runs immediately.

Also Read: AiThority Interview with Glenn Jocher, Founder & CEO, Ultralytics

This is the fifth free skill released by Speedrun AI Labs as part of its growing Skills Library, a curated collection of production-grade skills not regularly available in other skill marketplaces. Previous releases include /deep-audit (code security auditing), /no-slop-writing (AI writing quality control in American and Australian English editions), and /read-before-write (file safety for AI agents). All are available under MIT license on GitHub.

“We built this because we needed it ourselves,” Tanveer said. “We run two autonomous agents in production. When something better comes along, we do not rebuild them from scratch. We upgrade them with a skill file. Now any team can do the same.”

Speedrun AI Labs builds and deploys autonomous AI agents for Australian businesses through its ClawBot service. The company has published four research papers on AI agent architecture, two published and two pending, available for free download at speedrunlab.ai.

Also Read: ​​The Infrastructure War Behind the AI Boom

[To share your insights with us, please write to psen@itechseries.com]