Hermes Agent adds /learn command to distill reusable skills
TL;DR
Hermes Agent now supports the /learn command. The feature points at directories, URLs, or chat history to generate reusable skills that match project guidelines.
What changed
Developers can now use the /learn command in hermes-agent to teach new workflows automatically. Vibe builders point it at a directory, URL, or recent chat history to generate reusable skills. Basic users see these skills match their project's contributing guidelines without manual effort.
Why it matters
Developers maintain project standards more efficiently across repeated tasks. Vibe builders benefit in use-cases like multi-step code reviews where CrewAI requires separate configuration for each guideline set. Basic users apply distilled skills right away to keep outputs aligned.
What to watch for
Compare results against skills built in LangGraph on the same inputs. Verify by applying the generated skill to a fresh task from your chat history and confirming it follows the stated guidelines.
Who this matters for
- Vibe Builders: Point the /learn command at your best chat logs to turn one-off wins into reusable project skills.
- Basic Users: Use the /learn command to ensure your AI outputs automatically follow your team's style guidelines.
Harsh’s take
The /learn command in Hermes Agent addresses the biggest friction point in agentic workflows: the transition from a successful one-off prompt to a repeatable system. Most users struggle to formalize their 'vibe' into a structured prompt or tool definition. By allowing the agent to ingest directories or chat histories to distill skills, the barrier to entry for complex automation drops significantly.
This is a direct challenge to manual prompt engineering. Instead of writing long system instructions, you are essentially training the agent on your specific context via observation. It is a practical implementation of few-shot learning that actually lives where the work happens.
If the skill distillation is accurate, it makes maintaining project standards across a distributed team much easier than static documentation ever could.
by Harsh Desai
About Hermes Agent
View the full Hermes Agent page →All Hermes Agent updatesGo deeper
More AI news
- Daily RoundupVercel-Origin beta and GPT-5.6 Sol discount, plus agent tools for builders
Vercel rolled out Origin repository deployments and a month-long GPT-5.6 Sol discount on AI Gateway while Product Hunt featured new agent layers and Hugging Face highlighted an uncensored Qwen model, with industry moves around data use and automation exits.
- Weekly DigestHermes Agent v0.20 voice and A2A protocol, OpenClaw mobile updates, and multi-agent tools to test now
Hermes Agent rolled out v0.20 releases with voice streaming, A2A interoperability, desktop artifacts, and self-recovery tools while OpenClaw added Android switching, iOS caching, and plugin provenance checks across the week.
- Daily Roundupbeenga-image-1 on Replicate, Qwen3.8-27B on HF, and the AI trust debate
Fresh image and text models appeared on Replicate and Hugging Face while industry voices highlighted trust as the core barrier to wider AI adoption.