wquguru skills
A collection of practical skills: Claude 5 and GPT-5.6 model routing, English for engineers, Pi Agent setup, YouTube.
- 94 ★ on the repository
- Last commit
Minimal access
Why this level
- Instructions for the agent, part of the skills run helper scripts such as `sync_compat.py`
Install
- Works in
In your terminal, with SkillFoxx CLI
npx skillfoxx add skills/wquguru-skillsDetects the agents on your machine, shows the access level and pins the version.
Other ways to install
Access and checks
How we reviewMinimal accessThe tool runs nothing and makes no network calls.
- Instructions for the agent, part of the skills run helper scripts such as `sync_compat.py`
This is third-party code. Review the repository files before installing.
What it does
A monorepo of self-authored skills. `claude-5-best-practice` helps count the cost of an accepted result in the Claude 5 generation: when Opus 5 is enough, when Fable 5 is justified, how to raise effort and fix failures before switching models. `gpt-5-6-best-practice` does the same for GPT-5.6 in Codex. The other skills solve narrow tasks. `english-swe-daily` trains natural English for standups and code reviews, `pi-setup` configures Pi Agent with DeepSeek V4 and the Ring-2.6-1T model while working around seven known pitfalls, and `free-my-cpu` and `youtube-toolkit` cover system and video tasks. Each is installed with its own `npx skills add` command.
Who it is for. Engineers who pick models and effort levels for long tasks in Claude Code and Codex and work in English on a team.
Good fit when
- A run missed its acceptance criteria and you must decide whether to raise effort or change the model.
- Old prompts and skills are moving to Claude 5 and you need to find over-constrained parts and refusal risks.
- You want to sound natural and team-like in Slack and on calls.
- Pi Agent has to be connected to DeepSeek V4 with no keys in plaintext.
Not a fit when
- You need one specific skill: it installs from the monorepo with `--skill`, and installing everything is not advised.
- You need model advice for the long run: prices and limits in `references/evidence-notes.md` are dated and age.
- The task lies outside the set, for example a Pi project with neither DeepSeek nor Ring.
Example request
Does this task need Fable 5 or is Opus 5 at high effort enough?Limitations
The model routing skills rest on a dated snapshot of prices, limits and beta headers kept in `references/evidence-notes.md`, so after a while the facts must be checked against the source. Installation goes through `npx skills add` with the `--skill` flag and needs Node.js. `pi-setup` needs Pi Agent installed, and keys are read from your own file and never stored in plaintext. Third-party modules are declared in `external.yml` and pulled with `just vendor`. No license is declared in the repository.
How to disable. Remove the skill folder from the agent skills directory or uninstall through `npx skills`. Pi Agent settings written by `pi-setup` are rolled back by hand.
FAQ
How do I install a single skill?
With `npx skills add https://github.com/wquguru/skills --skill <name>`, for example `--skill pi-setup`.
Which agents do the skills work with?
The install links them into Claude Code, Codex and Grok skill directories, and the README calls the monorepo public and free of private tooling.
What is `claude-5-best-practice`?
A skill about the cost of an accepted result in Claude 5: effort levels, subagents with pinned models, API behavior and tying progress claims to real tool evidence.
What does `pi-setup` do?
It probes whether Pi Agent is installed and configured, then interactively connects DeepSeek V4, Ring-2.6-1T and an extension set without writing keys in plaintext.
Is there a security check?
Yes, the repository runs an NVIDIA SkillSpector audit and a Claude security review, with status badges in the README.
Discussion
No questions or reviews yet.
Related
Skills11 skills for Claude Code: SKILL.md authoring, prompt tuning, refactoring, change review and planning in files.
A collection of 14 portable skills for long agent runs: work planning, live app QA, search across past sessions and a phone call
A personal collection of about 130 Claude Code and Codex skills: media pipelines, Cloudflare, YouTube, design and task review
A Claude Code reference plus a collection of over 70 skills for development, security, data and docs, with agent configs
A working folder for Claude Code where the agent asks questions and writes a new skill by Anthropic's guidance, not from memory.
A skill for coding agents that pushes them to finish the job: keep trying after failures, switch approaches and prove the result
README in short
A list of self-authored monorepo skills with install commands, sample prompts and prerequisites. External modules and the GitHub Pages registry are described separately.