agentsclimarketplace

Format evaluator

Skill danielrosehill/Agent-Skills/skills/format-evaluator

Planning and "production" repo for agent skills

Install
npx -y skills add danielrosehill/Agent-Skills --skill format-evaluator

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.
  • 0 stars0 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

Evaluate whether a capability idea is best deployed as a Claude Code plugin, slash command, skill, subagent, or agent workspace. Use when the user describes an idea and asks "should this be a skill / slash command / plugin / agent?" or wants a recommendation on packaging format before building.

SKILL.md

2.3 KB, 502 tokens by cl100k_base, as published. Nobody here has run it

Format Evaluator

Decide the best packaging format for a Claude Code capability idea.

When to use

Invoke when the user:

  • Describes a workflow/idea and asks how to package it
  • Says things like "should this be a skill or a slash command", "is this a plugin", "where should this live"
  • Wants to convert an existing repo/workspace into a lighter format

Reference

Read docs/format-rubric.md in this repo for the full decision guide. The quick comparison table and the per-format "Use when / Don't use when / Signal phrases" sections are the primary decision aids.

Procedure

  1. Clarify the idea (if ambiguous): invocation trigger, state needs, who uses it, whether it ships to others.

  2. Apply the rubric — walk the five formats (Plugin, Slash Command, Skill, Subagent, Agent Workspace) and note fit/misfit signals from the rubric.

  3. Score: give each format a verdict — GOOD_FIT / PARTIAL / POOR_FIT — with one sentence of reasoning.

  4. Recommend: pick one primary format. If a hybrid makes sense (e.g. "slash command that invokes a skill", or "plugin bundling two skills + an agent"), state the bundle.

  5. Output in this shape:

    Idea: <one-line restatement>

Verdicts:
- Plugin: <GOOD/PARTIAL/POOR> — <reason>
- Slash Command: …
- Skill: …
- Subagent: …
- Agent Workspace: …

Recommendation: <format(s)>
Why: <2-3 sentences>
Next step: <concrete first action, e.g. "scaffold SKILL.md at skills/<name>/">

Heuristics cheatsheet

  • Model should reach for it automatically → Skill
  • User types /name to trigger → Slash Command
  • Heavy context / parallelizable side-quest → Subagent
  • Long-lived domain with accumulating files → Agent Workspace
  • Ships to others as a bundle → Plugin (wrapping the above)

Stateless + narrow + procedural = almost always a Skill. Stateful + ongoing + file-backed = almost always a Workspace.

What ships with it

Read from the repository

Just SKILL.md. No reference files, no scripts.

Keep looking

Skills are one crate of 327,069. 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.