Skip to content
Back to search

skills

Practical Agent Skills — English-for-engineers coaching, Pi Agent setup, and more. Install via npx skills add.

Safety scan

A static text scan, not a guarantee. Select the badge to see what was scanned and every finding.

Checked today
Badge for your README
findsafeskills safety rating

Markdown

[![findsafeskills safety rating](/api/listings/wquguru--skills/badge.svg)](/listings/wquguru--skills)

HTML

<a href="/listings/wquguru--skills"><img src="/api/listings/wquguru--skills/badge.svg" alt="findsafeskills safety rating" /></a>

Install

i

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.

README.md · README · 247 lines

View this file on GitHub (read at the repository's current version: the scan did not record which commit it read)

  1. # Skills
  2. [![NVIDIA SkillSpector audit](https://img.shields.io/github/actions/workflow/status/wquguru/skills/security-audit.yml?branch=main&logo=nvidia&label=NVIDIA%20SkillSpector%20audit)](https://github.com/wquguru/skills/actions/workflows/security-audit.yml)
  3. [![AI review by Claude](https://img.shields.io/badge/AI%20review-Claude-d97757?logo=claude)](https://github.com/anthropics/claude-code-security-review)
  4. A small collection of practical Agent Skills.
  5. > **Public monorepo.** Anyone can clone this repo, run `just add`, or install
  6. > individual skills with `npx skills add`. It does not depend on any private
  7. > tooling. The install loop links into Claude Code, Codex, and Grok skill dirs.
  8. **Catalog:** self-authored skills live under [`skills/`](./skills/) (see also the
  9. [GitHub Pages registry](https://wquguru.github.io/skills/)). Highlighted ones
  10. below; the full set is every folder with a `SKILL.md`. Third-party modules are
  11. declared in [`external.yml`](./external.yml) and materialised by `just vendor`.
  12. ## Skills
  13. ### `claude-5-best-practice`
  14. > Renamed from `fable-5-best-practice` in July 2026, when Claude Opus 5 shipped
  15. > near-Fable capability at half the price and the routing advice stopped being
  16. > about one model. Reinstall under the new name.
  17. Optimizes the cost of an *accepted result* across the Claude 5 generation —
  18. Haiku 4.5 / Sonnet 5 / Opus 5 / Fable 5 — rather than the price of a token. It
  19. treats Opus 5 as the default lead, makes Fable 5 something you justify instead
  20. of assume, and prefers climbing the effort ladder (`low` → `max`) on Opus 5
  21. before escalating a tier. It also walks a diagnose-before-escalating loop for
  22. failures, routes subagents with pinned worker models, covers Claude 5 API
  23. behavior (adaptive thinking, effort and prompt caching, refusal stop reasons,
  24. fallback, ZDR and Priority Tier constraints), and keeps progress claims tied to
  25. real tool evidence. Volatile facts such as pricing, retention, and beta headers
  26. live in a dated `references/evidence-notes.md` provenance snapshot.
  27. The skill is especially useful when migrating older prompts or skills forward:
  28. it calls out over-constrained legacy prompting, the Opus 4.8 → Opus 5 thinking
  29. and effort breaks, refusal risks, verifier agents, long-running run design, and
  30. long-run communication style.
  31. Install:
  32. ```bash
  33. npx skills add https://github.com/wquguru/skills --skill claude-5-best-practice
  34. ```
  35. Ask for help with prompts like:
  36. - "Does this task actually need Fable 5, or is Opus 5 enough?"
  37. - "Help me write an Opus 5 prompt for a multi-day coding task."
  38. - "Review this old skill before migrating to Claude 5."
  39. - "My run missed the acceptance criteria — raise effort or change model?"
  40. ### `gpt-5-6-best-practice`
  41. A routing and evaluation overlay for GPT-5.6 in Codex. It optimizes the cost of
  42. an accepted result — not the price of a token — across the Sol, Terra, and Luna
  43. capability tiers: choosing the initial tier and reasoning-effort lane,
  44. diagnosing failures before escalating, keeping prompts and worker packets lean,
  45. using subagents economically, and evaluating routes on real work. Benchmark
  46. provenance and confidence levels live in `references/evidence-notes.md`, and
  47. `references/prompt-patterns.md` provides reusable task-prompt and worker-packet
  48. templates.
  49. Install:
  50. ```bash
  51. npx skills add https://github.com/wquguru/skills --skill gpt-5-6-best-practice
  52. ```
  53. Ask for help with prompts like:
  54. - "Which GPT-5.6 tier and effort should this task start on?"
  55. - "This Codex run failed — raise effort, or move from Terra to Sol?"
  56. - "Is Ultra worth it for this task, or should one agent handle it?"
  57. - "Design a worker packet for this subagent."
  58. ### `english-swe-daily`
  59. A practical English coach for everyday engineering communication. Built for
  60. developers who can already work in English but want to sound less textbook-like
  61. and more like a real teammate in standups, Slack threads, 1:1s, code reviews,
  62. and meetings.
  63. Many engineers do not struggle with grammar — they struggle with sounding
  64. natural. This skill teaches the kind of English people actually use in software
  65. teams, with a strong focus on natural phrasing, collaboration tone, and
  66. day-to-day workplace situations.
  67. Install:
  68. ```bash
  69. npx skills add https://github.com/wquguru/skills --skill english-swe-daily
  70. ```
  71. Ask for help with prompts like:
  72. - "Teach me English for standups."
  73. - "Help me sound more natural in Slack."
  74. - "Practice English for code reviews."
  75. - "How do I say this more naturally at work?"
  76. - "Give me today's SWE English session."
  77. ### `pi-setup`
  78. Configures Pi Agent (`@earendil-works/pi-coding-agent`) with DeepSeek V4
  79. (built-in provider) and Ant-Ling Ring-2.6-1T (custom OpenAI-compatible
  80. provider), plus a curated extension set, while automatically working around 7
  81. known pitfalls.
  82. It detects before it acts: it probes whether pi is installed, whether config
  83. already exists, and where keys live before changing anything. Keys are never
  84. written in plaintext — pi config only stores lazy `!shell` commands and the key
  85. stays in the user's original file. Model selection, default model, default
  86. thinking tier, and the extension set are all confirmed interactively.
  87. Install:
  88. ```bash
  89. npx skills add https://github.com/wquguru/skills --skill pi-setup
  90. ```
  91. Trigger it in Claude Code by saying "配置 pi", "setup pi agent", "把
  92. ring/deepseek 配到 pi", or run `/pi-setup` manually.
  93. Prerequisites:
  94. - Node.js / npm (to install pi)
  95. - A DeepSeek key (in `~/.deepseek` or an environment variable)
  96. - An Ant-Ling/Ling key (`LING_API_KEY`, in a shell rc or environment
  97. variable) — only needed when enabling Ring
  98. ### `free-my-cpu`
  99. Safely audits CPU, load average, I/O wait, and Docker container pressure on
  100. Linux hosts, especially production-like servers over SSH. It follows the same
  101. read-only-first posture as `free-my-disk`: collect repeated samples, separate
  102. CPU-bound work from I/O pressure, inspect hot containers and logs, then classify
  103. possible fixes by impact before changing anything.
  104. Install:
  105. ```bash
  106. npx skills add https://github.com/wquguru/skills --skill free-my-cpu
  107. ```
  108. Ask for help with prompts like:
  109. - "Use free-my-cpu to inspect this remote Docker host."
  110. - "Why is this Docker host load average so high?"
  111. - "Find which container is burning CPU."
  112. - "Use this Grafana URL and the last 6 hours to check whether monitoring queries are causing load."
  113. ### `youtube-toolkit`
  114. Downloads YouTube videos with `yt-dlp` and post-processes them with `ffmpeg`:
  115. fetch best-quality streams, convert containers (e.g. AV1/webm → H.264 mp4),
  116. extract audio, and burn in translated subtitles with a translator watermark.
  117. It is macOS-first and bakes in the gotchas learned the hard way — chiefly that
  118. Homebrew's lean `ffmpeg` ships without libass/drawtext (so subtitle burn-in
  119. needs `ffmpeg-full`), that YouTube auto-captions are overlapping rolling cues
  120. that must be re-segmented before they read cleanly, and that styling via an ASS
  121. file avoids the `force_style` comma-parsing trap. Two helper scripts do the
  122. deterministic work: `install-tools.sh` and `burn-subs.sh`.
  123. Install:
  124. ```bash
  125. npx skills add https://github.com/wquguru/skills --skill youtube-toolkit
  126. ```
  127. Ask for help with prompts like:
  128. - "Download this YouTube video to ~/Downloads."
  129. - "Convert it to mp4."
  130. - "Add Chinese subtitles and a @handle 翻译 watermark."
  131. Prerequisites:
  132. - Homebrew (macOS) or apt/pip (Linux)
  133. - `yt-dlp` and a libass-enabled `ffmpeg` — `install-tools.sh` sets both up
  134. ## Local Development
  135. For local hacking, the `Justfile` is **self-contained**: it symlinks every skill
  136. under `skills/` **and** `external/` into Claude Code (`~/.claude/skills/`),
  137. Codex (`~/.codex/skills/`), and Grok (`~/.grok/skills/`), so edits take effect
  138. immediately without re-publishing. No private sibling repo is required.
  139. ```bash
  140. just add # symlink all skills (skills/ + external/) into agent dirs
  141. just remove # remove only the symlinks that point back into this repo
  142. just status # show install status for each skill in all destinations
  143. just vendor # pull third-party skills declared in external.yml into external/
  144. just # list available recipes
  145. ```
  146. `add` is idempotent and never clobbers: an existing real directory or a symlink
  147. from another install tree is left untouched and reported as skipped
  148. (`foreign symlink`). `remove` only removes symlinks that point back into this
  149. repo. After adding a new skill folder, re-run `just add`.
  150. ## External (third-party) skills
  151. `skills/` holds the skills authored here (and published via `npx skills add`).
  152. Third-party skills from other repos are **not** copied into git — they are
  153. vendored separately so this public repo stays purely self-written and free of
  154. foreign licenses:
  155. 1. Declare sources in `external.yml` — a repo, a `ref` (branch/tag/commit), and
  156. the sub-paths to pull. Each dir's basename, prefixed with `3rd-`, becomes the
  157. skill name (e.g. `dashboarding` → `3rd-dashboarding`).
  158. 2. `just vendor` shallow/sparse-clones each source, copies the named skill dirs
  159. into `external/` (git-ignored) under their prefixed names, rewrites each
  160. `SKILL.md` frontmatter `name` to match, and records the exact upstream commit
  161. SHA in `external.lock`.
  162. 3. `just add` symlinks them alongside the self-written skills.
  163. Only `external.yml` + `external.lock` are committed; the vendored code is not.
  164. Re-run `just vendor` any time to refresh (or after pinning a `ref`). A name that
  165. would collide with a self-written `skills/<name>` is skipped; entries removed
  166. from `external.yml` are pruned from `external/` on the next `just vendor`.
  167. ## Repository Structure
  168. ```text
  169. Justfile # add / remove / status / vendor (self-contained)
  170. external.yml # third-party skill sources (repo + ref + sub-paths)
  171. external.lock # resolved upstream commit SHAs (auto-generated)
  172. scripts/
  173. vendor.sh # pulls external.yml sources into external/
  174. skills/ # self-written skills (published)
  175. claude-5-best-practice/
  176. gpt-5-6-best-practice/
  177. english-swe-daily/
  178. …
  179. external/ # vendored third-party skills (git-ignored; just vendor)
  180. index.html # public site
  181. ```
  182. ## License
  183. Self-authored skills under `skills/` are intended for public use via this repo
  184. and `npx skills add`. Third-party code under `external/` is **not** redistributed
  185. in git — only `external.yml` / `external.lock` are tracked; run `just vendor` to
  186. fetch upstream under their own licenses.

Ratings and comments

Sign in to rate this skill.

Sign in

Free to use. If it’s helped, pass it on.