Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
npx skills add shihyuho/pangu.skill --skill pangugit clone --depth 1 https://github.com/shihyuho/pangu.skillWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/skills/shihyuho/pangu.skill/pangu)<a href="https://agentmods.dev/skills/shihyuho/pangu.skill/pangu"><img src="https://agentmods.dev/badge/skills/shihyuho/pangu.skill/pangu/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/skills/shihyuho/pangu.skill/pangu"><img src="https://agentmods.dev/badge/skills/shihyuho/pangu.skill/pangu.svg" alt="Reviewed on agentmods" width="80" height="20"></a>What it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.00068 | $0.01705 |
| Opus 5 | $0.00034 | $0.00852 |
| Sonnet 5 | $0.00014 | $0.00341 |
| Haiku 4.5 | $0.00007 | $0.00170 |
Grade A, and why
pangu scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured today.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 135 lines — stays where its author put it; the contents beside it link to each section on GitHub.
pangu
Paranoid Text Spacing (盤古之白)
Scope first
- Follow the user's requested format and editing scope. Space the prose you create or modify for the task; format an entire document when that is the requested task. Keep unrelated text as received.
- Preserve literal content byte-for-byte: code, inline code, fenced or indented code blocks, exact quotations, command output, and logs, even outside code spans. Change a literal only when the user's task explicitly calls for it.
- Apply spacing to natural-language wording, including headings, link labels, and code comments. Preserve Markdown markers, link destinations, and other structural syntax, plus the interiors of URLs, paths, email addresses, versions, and compound identifiers. A Chinese character inside a token is still part of that token.
- Deliver the requested work with spacing already applied.
For example, space a Markdown link's visible label while keeping its URL exact. When quoting a log for searching, preserve the log and space only your own explanation. These boundaries take precedence over the text rules below.
Text rules
Insert one half-width space where CJK meets half-width letters or numbers;
use the symbol rules below for punctuation and operators. CJK includes Han
ideographs, Japanese kana, and bopomofo, including Han used in Korean.
Hangul is outside this definition: 한국어test stays unchanged.
Full-width punctuation (，。！？「」) needs no surrounding space.
For spacing, leave runs without CJK and already-correct whitespace unchanged.
當你凝視著bug，bug也凝視著你 → 當你凝視著 bug，bug 也凝視著你
與PM戰鬥的人 → 與 PM 戰鬥的人
這是2025年的事 → 這是 2025 年的事
Operators and separators
+ - * = ^ < > \: space CJK boundaries, keeping half-width words whole. Before CJK, also separate- * = &from a preceding) ] }. A hyphen directly between CJK and digits is a separator; an already-spaced negative number stays attached to its sign._: keep attached in every context.|: if one pipe touches CJK, space every pipe on that line; otherwise preserve half-widthA|B.- Single
/and&: keep half-width joiners (A/B,R&D) whole; space them when CJK is adjacent. Two or more slashes keep their path structure.
What ships with it
1 file beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- today Changed · -94 lines · -56 tokens per session 0b3096d4cd1a
- 10d ago First seen · 229 lines · 124 tokens per session scan A 5ab2ba6474cd
pangu is a skill published in the GitHub repository shihyuho/pangu.skill (2 stars, last pushed today), licensed MIT. It adds 68 tokens to every session and 1,705 once invoked, about $0.0003 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.
Other skills, from other repositories
lov-human-writing
A Chinese-language editing workflow for making drafts sound more like their author while preserving the author's judgments and uncertainty.
lov-image-translation-errata
A workflow for correcting machine-translated text inside screenshots or other images while keeping the original wording, mistakes, layout, and visible correction marks.
lov-translation-review
A review tool for comparing a Chinese source document with its English translation. It checks meaning, grammar, terminology, and consistency, then produces a prioritized report.
lov-academic-translator
A tool for translating English research papers, journal or conference articles, arXiv papers, and technical PDFs into Chinese PDFs or publishable Chinese Markdown. It can preserve requested images, formulas, page matching, and layout details.
lov-subtitle-freedom
Create learner-friendly English subtitles with level-aware glosses and optional spoiler-safe subtitle sidecars.
lov-add-tool
A workflow for adding a working online tool to an existing website. It covers the tool page, its listing, translations, site map entry, input handling, results, and cost notices when paid services are used.