Skip to the rows

agentsclimarketplace

Skill class model en

Skill roebi/new-work-skills/skills/skill-class-model-en

Agent Skills that teach and implement the **New Work with Skills** methodology

Install

npx -y skills add roebi/new-work-skills --skill skill-class-model-en

Assembled from the repository path, not quoted from the project. Check it against their README if it does not work.

2 things to look at

no licenseNo license file was found in the repository. Code published without one is not open source by default, so using it at work is a question for whoever answers licensing questions where you are.
2 stars2 stars. Stars are a popularity signal and not a quality one, but at this level it is likely that nobody has read this closely except its author, and you would be relying on your own review.

What its author says it does

Copied from the file, not written here

Documents the Skill Class Model — the OO-based taxonomy that classifies every Agent Skill as one of: Role, Topic, Process, or OneStepProcess, each extending the base Skill class. Defines the v1 metadata extension fields for skill-class, crud-verb, topic, requires-role, subprocess-of, subtopic-of, and subrole-of. Use when an agent or author needs to classify a skill, assign metadata, understand the class hierarchy, or check naming conventions. Activate for phrases like: "what class is this skill", "how do I classify a skill", "skill taxonomy", "skill-class metadata", "is-a or has-a for skills", "what metadata does a Role need", or "naming convention for process skills".

The file declares its own license as CC BY-NC-SA 4.0. That is the author’s claim about this one file, and it is not the same thing as the license GitHub reports for the repository, which is listed with the other numbers below.

SKILL.md

3.8 KB, 724 tokens by cl100k_base, as published. Nobody here has run it

Skill Class Model

The Skill Class Model applies object-oriented design to Agent Skills. Every skill is an instance of a class that extends the base Skill class. Classes are discriminated by the metadata.skill-class field (v1 extension).

Class Hierarchy (is-a / inheritance)

Skill                          ← base class (agentskills.io)
├── Role                       ← extends Skill
├── Topic                      ← extends Skill
└── Process                    ← extends Skill
    └── OneStepProcess         ← extends Process

Composition (has-a)

Topic          has  Subtopic        1:n
Topic          has  File            1:n

Role           has  Subrole         1:n

Process        needs  Role          1:n   (mandatory)
Process        has    Subprocess    1:n

OneStepProcess needs  Role          0..1  (optional)

Key rule: Process is higher than Role. A Process needs one or more Roles — never the other way around.

The Subject and the Verb

A Topic is the subject — the thing that gets operated on. A OneStepProcess applies a CRUD verb to a Topic.

crud	natural language equivalents
create	create, generate, scaffold, write, init
read	read, search, find, list, get, describe, show
update	update, fix, move, rename, migrate, refactor
delete	delete, remove, clean, prune

Naming Conventions

Class	Pattern	Example
Role	`role-<name>-en`	`role-sw-architect-en`
Topic	`<topic>-en` or `<compound-topic>-en`	`agentic-runtime-architecture-en`
Process	`<name>-en`	`onboarding-employee-en`
OneStepProcess	`<crud-verb>-<topic>-en`	`create-agent-skill-en`

Metadata Fields — v1

See references/metadata-v1.md for the full field reference.

Quick summary:

# All classes — discriminator (required for classified skills)
metadata:
  skill-class: role | topic | process | one-step-process

# Role adds:
  subrole-of: <role-name>           # optional

# Topic adds:
  subtopic-of: <topic-name>         # optional
  requires-role: <role-name>        # optional

# Process adds:
  requires-role: <role-name>        # 1:n comma-separated (mandatory)
  subprocess-of: <process-name>     # optional

# OneStepProcess adds:
  crud-verb: create | read | update | delete   # mandatory
  topic: <topic-name>                          # mandatory
  subprocess-of: <process-name>               # optional
  requires-role: <role-name>                  # optional 0..1

Composability Chain

When an agent loads a skill it resolves the role chain automatically:

load skill
  → read requires-role
    → load role-<name>-en
      → agent has role context + task context
        → execute

References

references/metadata-v1.md — full field-by-field metadata reference
references/examples.md — concrete skills mapped to each class

What ships with it: 2 files

5.1 KB alongside SKILL.md

references/

examples.md1.3 KB
metadata-v1.md3.8 KB

Keep looking

Skills are one crate of 328,083. Ordering is by how many stacks a row turns up in, so the top of any crate is what has actually been picked rather than what has the most stars.