Passed 19 of 22 test cases with the skill, and 9 of 22 without it.
Without skill41%
With skill86%
Checks it passed that the model alone missed
State that the assistant is named Hermes.
Clarify that the assistant can read local files.
Identify the FLUX model as the default image model.
and 13 more in the full report
10 fixed9 passed either way3 still failing0 broke
Pass-rate lift
+45 pp
Tokens
+25%
Agent turns
0%
Response time
-44%
Model gemini-3.6-flashMeasured 10/10/2026v119 of 22 cases passed with skill
Tokens, turns and time describe resource use, not a dollar price.
See it in action
One of 10 test cases the skill fixed. Same model, same prompt, with and without the skill.
1 of 10
Prompt
Where in the Portable Hermes desktop app interface do I check and manage available extensions like ComfyUI or TTS?
What was checked, without and with the skill
Checked
Without skill
With skill
Direct the user to check the View menu under Extensions.
Missed
Passed
case-11 · gemini-3.6-flash · 10/10/2026 · answers are capped and redacted
What's inside
The instructions your model reads when the skill loads.
SKILL.md121 linessingle file
Getting Started with Hermes Agent
When a user asks for help, what you can do, or seems new — use this guide to give them a warm, friendly overview. Do NOT dump this entire document. Pick the parts that match what they're asking about and explain in your own words.
When to Use
Use this skill when a user is new to Portable Hermes or asks about its setup, GUI, included tools, supported tasks, or optional extensions.
For brand new users, explain it like this:
"I'm Hermes, your AI assistant. Think of me as a really capable helper that lives on your computer. I can browse the web, create files, generate images and music, read documents, and remember things between our chats. Just talk to me like you'd talk to a friend — ask me anything and I'll figure out how to help."
What you can do — organized by what users actually want
Finding Information
Search the web — "What's the weather in Tokyo?" / "Find me a good pasta recipe"
Read websites — "Summarize this article: [url]" / "What does this website say?"
Browse interactively — You can open web pages, click buttons, fill forms, take screenshots
Example prompts: "Search for...", "Look up...", "What is...", "Find me..."
Working with Files
Read files — "What's in this document?" / "Read my notes.txt"
Write files — "Create a shopping list and save it" / "Write a letter for me"
Edit files — "Fix the typo in my file" / "Add a line to my list"
Find files — "Find all .txt files in my documents"
Example prompts: "Save this as...", "Create a file...", "What's in..."
Creating Images
Generate images from descriptions — "Draw me a sunset over mountains" / "Create a logo for my bakery"
Images are generated by AI (FLUX model via FAL.ai)
You can be as descriptive as you want — style, colors, mood, composition
Example prompts: "Generate an image of...", "Create a picture of...", "Draw me..."
Music & Audio
Generate music (requires Music Server extension) — "Make me a chill lo-fi beat" / "Create an epic orchestral piece"
Text-to-speech — "Read this text out loud" / "Say 'Happy Birthday' in a cheerful voice"