New agent

Skill hcussi/claude-code-toolkit/.claude/skills/new-agent

Scaffold a new agent in this claude-code-toolkit repo (create agents/<name>.md, the mirrored docs/agents/<name>.md, and add the README table row) following repo conventions. Use when adding a new agent to the toolkit.From its SKILL.md

Install

npx -y skills add hcussi/claude-code-toolkit --skill new-agent

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.
1 stars1 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.

SKILL.md

2.1 KB, 507 tokens by cl100k_base, as published. Nobody here has run it

new-agent

Scaffold a new toolkit agent so every entry lands consistent: the definition, its mirrored doc, and the README table row, in one pass.

Inputs

Ask the user for these if not provided:

name: kebab-case, unique (becomes agents/<name>.md). Verify it does not already exist under agents/.
description: one or two sentences covering what it does and when to use it. This drives automatic delegation, so make it specific.
tools: the minimal set the agent needs (e.g. Read, Grep, Glob, Bash for a read-only reviewer). Omit to inherit all tools.
model: optional (sonnet, opus, haiku); omit to inherit.

Steps

Create agents/<name>.md with this frontmatter, then the system prompt body:

---
name: <name>
description: <description>
tools: <comma-separated tools>   # omit line to inherit all
model: <model>                   # omit line to inherit
---

<System prompt: role, scope, method, and reporting/output format.>

Create docs/agents/<name>.md: the human-facing doc covering purpose, scope, how to use it, and ideally a short example of its output. Go beyond the frontmatter; a one-line stub will fail review.

Add a README row. In the ## Agents table in README.md, add:

| `<name>` | <short description> | [docs](docs/agents/<name>.md) |

Validate. Run .claude/scripts/lint-toolkit.sh and fix anything it reports. For a qualitative pass, suggest the user invoke the toolkit-linter agent.

Notes

Keep the top-level agents//skills/ folders for published toolkit items only. Repo-maintenance tooling lives in .claude/.
Follow repo conventions: no em dashes; single root .gitignore.
The pre-commit hook runs the lint script automatically, so a scaffold that skips step 2 or 3 will block the commit until fixed.

What ships with it

Read from the repository

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