Back to search
theamazingwolf
MarketplaceWorks with Claude CodeGitHub
Public agent skills and Claude Code plugins by Tim Wolf.
Safety scan
A static text scan, not a guarantee. Select the badge to see what was scanned and every finding.
Badge for your README
Markdown
[](/listings/theamazingwolf--skills)HTML
<a href="/listings/theamazingwolf--skills"><img src="/api/listings/theamazingwolf--skills/badge.svg" alt="findsafeskills safety rating" /></a>Install
Install instructions from the README
As written by the repo's author. findsafeskills has not verified these steps.
Any agent via `npx skills`
npx skills add theamazingwolf/skills
List the skills without installing:
npx skills add theamazingwolf/skills --list
Install one skill only:
npx skills add theamazingwolf/skills --skill wlf-jev
Claude Code
/plugin marketplace add theamazingwolf/skills /plugin install wlf-skills@theamazingwolf
From your shell: `claude plugin marketplace add theamazingwolf/skills`, then `claude plugin install wlf-skills@theamazingwolf`.
Codex
codex plugin marketplace add theamazingwolf/skills codex plugin add wlf-skills@theamazingwolf
GitHub Copilot CLI
copilot plugin marketplace add theamazingwolf/skills copilot plugin install wlf-skills@theamazingwolf
Pick one route per agent. Installing the same skill through both `npx skills` and the plugin loads it twice.
What was scanned
The files the safety scan read, as written by the repository's author and shown as plain text. Lines with a finding are marked.
.claude-plugin/marketplace.json · Manifest · 32 lines
- {
- "name": "theamazingwolf",
- "description": "Public agent skills and Claude Code plugins by Tim Wolf.",
- "owner": {
- "name": "Tim Wolf",
- "url": "https://timwolf.com"
- },
- "metadata": {
- "description": "Public agent skills and Claude Code plugins by Tim Wolf.",
- "version": "1.0.0"
- },
- "plugins": [
- {
- "name": "wlf-skills",
- "source": "./plugins/wlf-skills",
- "description": "Tim Wolf's public agent skills. Includes wlf-jev: design, code, calibrate and run TypeSafe Jev typed-judgment questions in your codebase, with runner and calibration scripts, TypeScript starters and a visual explainer.",
- "version": "1.0.0",
- "author": {
- "name": "Tim Wolf"
- },
- "license": "MIT",
- "homepage": "https://github.com/theamazingwolf/skills",
- "category": "development",
- "keywords": [
- "agent-skills",
- "typesafe",
- "jev",
- "calibration"
- ]
- }
- ]
- }
README.md · README · 122 lines
- # Skills
- Agent skills by Tim Wolf · [timwolf.com](https://timwolf.com) · GitHub [@theamazingwolf](https://github.com/theamazingwolf) · X/IG @realtimwolf
- Each skill is a self-contained folder under `plugins/wlf-skills/skills/` (`SKILL.md` plus its
- references, scripts and assets) that works in any agent that reads skills.
- The whole pack also installs as one plugin, `wlf-skills`, which adds slash
- commands as entry points. Install the pack once; each skill below lists its
- own requirements.
- ## Install
- ### Any agent via `npx skills`
- ```bash
- npx skills add theamazingwolf/skills
- ```
- List the skills without installing:
- ```bash
- npx skills add theamazingwolf/skills --list
- ```
- Install one skill only:
- ```bash
- npx skills add theamazingwolf/skills --skill wlf-jev
- ```
- ### Claude Code
- ```
- /plugin marketplace add theamazingwolf/skills
- /plugin install wlf-skills@theamazingwolf
- ```
- From your shell: `claude plugin marketplace add theamazingwolf/skills`,
- then `claude plugin install wlf-skills@theamazingwolf`.
- ### Codex
- ```bash
- codex plugin marketplace add theamazingwolf/skills
- codex plugin add wlf-skills@theamazingwolf
- ```
- ### GitHub Copilot CLI
- ```bash
- copilot plugin marketplace add theamazingwolf/skills
- copilot plugin install wlf-skills@theamazingwolf
- ```
- Pick one route per agent. Installing the same skill through both
- `npx skills` and the plugin loads it twice.
- ## Skills
- | Skill | What it does |
- |---|---|
- | [wlf-jev](plugins/wlf-skills/skills/wlf-jev/) | Design, code, calibrate and run TypeSafe Jev typed-judgment questions in your codebase |
- ### wlf-jev
- #### What it does
- Turns "classify / check / score / route X" into TypeSafe Jev questions,
- working TypeScript in your repo, and a calibration report that says how far
- to trust them. It picks the primitive (noul, choice, score, box), writes the
- wording and state, measures agreement against a small set of human labels,
- and wires the result in as an agent tool, in-app classifier, agent guard
- check or batch run. Ships a calibration script, a question runner with an
- offline mock mode, TypeScript starters and a visual explainer.
- The rule it follows: Jev makes judgments and code makes decisions, and no
- threshold ships without agreement measured against human labels.
- #### What you need
- - Node 22 or newer.
- - A [TypeSafe](https://typesafe.ai) account and an API key in
- `TYPESAFE_API_KEY` (environment or a git-ignored `.env`). Without a key,
- the question runner's `--mock` mode still supports offline design and
- pipeline tests.
- - **Data notice:** whatever goes into a Jev question's `state` (and its
- `instructions` or `criteria`) is sent to TypeSafe, a third-party service.
- Keep secrets out, and check that the data's owner is fine with that.
- #### Commands
- Plugin commands (Claude Code):
- | Command | Skill mode |
- |---|---|
- | `/wlf-skills:jev-setup [client-project-path]` | Check Node, SDK and key; make a smoke call |
- | `/wlf-skills:jev-explain [question]` | Open the visual explainer |
- | `/wlf-skills:jev-design <what-to-judge>` | Pick primitives, write questions, state and a confidence plan |
- | `/wlf-skills:jev-calibrate [spec] [labelled-items]` | Measure agreement against human labels, pick thresholds |
- | `/wlf-skills:jev-integrate [shape] [spec]` | Wire in as agent tool, classifier, guard check or batch run |
- | `/wlf-skills:jev-run <spec> <data>` | Estimate cost, then run at scale |
- Or just ask ("add a Jev check for...") and the skill triggers on its own.
- _Not affiliated with or endorsed by TypeSafe. TypeSafe and Jev are trademarks of their respective owners._
- ## Layout
- ```
- .claude-plugin/marketplace.json marketplace "theamazingwolf"
- plugins/wlf-skills/
- .claude-plugin/plugin.json plugin "wlf-skills" (bundles every skill)
- commands/ slash commands (plugin installs only)
- skills/<skill>/ each skill: SKILL.md, references, scripts, assets
- ```
- `npx skills` finds every skill under `plugins/wlf-skills/skills/` through
- the marketplace file, so each one is installable on its own.
- ## License
- [MIT](LICENSE) © 2026 Tim Wolf
Ratings and comments
Sign in to rate this skill.
Sign inFree to use. If it’s helped, pass it on.