AI Support Agent Skill Library

Library

Idemeum skill library contains 14 skills and 107 tools.

Skills

Skills are instructions written in Markdown that teach the AI Support Agent how to resolve specific endpoint issues. For example, disk-cleanup frees up storage, and process-manager diagnoses slow computers. Each skill defines when to use it, what steps to follow, and which tools it can invoke. Skills contain:

Section	Description
Frontmatter	Top YAML block that declares structured metadata like the skill’s name, risk level, allowed tools. etc.
Body	Plain Markdown with the agent’s playbook: when to use the skill, the steps to take, edge cases, etc.

Tools

Tools are executable capabilities the AI agent can invoke from inside the skill. These are narrowly scoped deterministic functions like kill_process or list_printers that perform one well-defined action and returns a structured result. Tools contain:

Section	Description
Meta block	Declares structured metadata such as the tool’s name, description, risk level, whether it requires user consent, etc.
Body	Actual code that runs when the agent calls the tool, including any platform-specific branches (macOS vs Windows), error handling, and the shape of the data it returns.