Agent Skills (magnus919)
magnus919/agent-skills
Curated governance, evaluation and operating-economics skills for Hermes and other agents
This repository is a curated collection of reusable AI agent skills, workflows, protocols and knowledge packs, following the open Agent Skills format, that codify patterns for making agents more capable, governable and trustworthy, for Hermes and other agent frameworks.
What Agent Skills (magnus919) does
Rather than general-purpose prompts, skills here cover specific operating disciplines: evaluations and production operations, AI governance, operating economics, and multi-agent collaboration, alongside domain skills like actuarial risk modeling, architecture decision records and multi-agent structured debate. Each skill is a self-contained folder with a SKILL.md, following the Agent Skills open format so it works in any framework that supports the standard.
Several skills are more than reference text: agent-council, for example, ships as a pip-installable package built on PydanticAI and PydanticGraph that spawns a panel of expert agents to debate a question with convergence-aware iteration, producing a decision landscape with confidence diagnostics and a principal's recommended path. The collection is updated frequently and documents its own meta-skill for creating or reviewing any other skill in the repository.
Key features
- Governance, evaluation and operating-economics skills alongside domain-specific ones
- Standard Agent Skills format, usable in any compliant framework including Hermes
- agent-council: a pip-installable multi-agent debate package with confidence diagnostics
- A meta-skill documenting how to create or review any other skill in the collection
- Full-catalog install plus an update command to keep installed skills current
When to use it
- Adding production-operations and governance discipline to an agent deployment
- Running a structured multi-agent debate on a decision with agent-council
- Using the actuarial-risk-modeling or adr-authoring skills for a specific domain task
Who it is for: Teams operating AI agents in production who want governance, evaluation and multi-agent collaboration patterns as installable skills, not just internal documentation.
How it fits with Hermes Agent
The repository describes itself as a collection of skills for Hermes and other agent frameworks, built on the framework-neutral Agent Skills open format.
FAQ
What is Agent Skills (magnus919)?
It is a curated set of Agent Skills covering AI governance, evaluations, operating economics and multi-agent collaboration, following the open Agent Skills format so any compliant agent, including Hermes, can use them.
Does it work with Hermes Agent?
Yes, the project's own description names Hermes as one of the agent frameworks the skills are built for, alongside other frameworks that support the Agent Skills standard.
Is this collection free and open source?
Yes, the repository is released under the MIT license.
Similar skills for Hermes Agent
All skillsCurated, field-tested skills for Hermes Agent covering diagnostics, audits, handoffs and Kanban planning
Da7-Tech SureForgeInstruction-only skill that makes an agent research and verify before calling work done
Hmbown Wizards of the Ghosts123 Hermes skills built from fantasy spell names, for investigation and repair
hezkvectory Hermes Edu Skills170 Chinese-education skills for textbook sync, exam prep and family learning
gongdear cline-pilotAgent Skill that drives Cline CLI coding tasks for you and learns your per-project decision preferences
Forlives 21 Days of Self-InterviewHermes Agent skill that asks three questions each night for 21 nights and remembers your answers
Related guides: What is Hermes Agent? · SOUL.md for Hermes Agent: what it is and how to write one