Getting it into your agent
It runs from inside its repository, so the clone comes first — what it calls does not travel with the file alone.
git clone --depth 1 https://github.com/rojim666/SztuCodenpx agentmods add skills/rojim666/sztucode/html-reviewWrote 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/rojim666/sztucode/html-review)<a href="https://agentmods.dev/skills/rojim666/sztucode/html-review"><img src="https://agentmods.dev/badge/skills/rojim666/sztucode/html-review/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/rojim666/sztucode/html-review"><img src="https://agentmods.dev/badge/skills/rojim666/sztucode/html-review.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 | $0.00165 | $0.01698 |
| Opus 5 | $0.00082 | $0.00849 |
| Sonnet 5 | $0.00033 | $0.00340 |
| Haiku 4.5 | $0.00016 | $0.00170 |
Grade A, and why
html-review 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 — 131 lines — stays where its author put it; the contents beside it link to each section on GitHub.
html-review Skill
何时使用本 Skill
在 doc-formatter 的路线 A 和路线 B 流水线中,每次 HTML 输出后调用本 skill 执行质量门禁:
- 路线 A:每次
doc-typeset输出 HTML 后调用 - 路线 B:每次
doc-edit输出 HTML 后调用
若检测不通过,将修正反馈返回对应 skill(路线 A → doc-typeset;路线 B → doc-edit)做一次定向修正,修正后直接输出,不再复检、不循环。
职责
执行 6 维度 HTML 质量门禁检测,输出结构化 HtmlReviewReport。安全性维度(SC-01~SC-06)为一票否决维度;其余 5 维度参与综合分计算。检测不通过时,生成可操作的修正建议返回给上游 skill。
执行方式(脚本化,0 次 LLM 往返)
6 个维度的检测项全部为确定性规则(正则匹配、标签计数、结构解析、class/属性存在性判断),已由 scripts/review_html.py 完整实现(纯 Python 3 标准库,零第三方依赖)。本 skill 直接运行脚本并解析其 JSON 输出作为 HtmlReviewReport,无需读取 references/ 逐条人工检测,无需任何 LLM 推理往返。
python3 scripts/review_html.py --html <html_path> --genre <genre>
# 或从 stdin 读取:
cat <html_path> | python3 scripts/review_html.py --stdin --genre <genre>
- stdout 输出即为完整
HtmlReviewReport(JSON)。 - 退出码:
0= 通过(passed=true);1= 不通过(passed=false);2= 运行错误(输入缺失/为空)。 - 脚本失败(退出码 2 或异常)→ 视为
review_skill_failed,由 doc-formatter 输出当前最佳 HTML(见 doc-formatter §7)。
references/ 目录下的规则文件是脚本实现的规则来源与说明,仅供维护脚本时对照,运行时无需加载。
输入契约
interface HtmlReviewInput {
html: string; // doc-typeset 输出的完整 HTML 字符串
design_tokens: object; // 与排版时相同的 design token 对象
genre: string; // 文档类型(government-doc / legal-contract / academic-paper 等)
}
输出契约
interface HtmlReviewReport {
passed: boolean; // 整体是否通过(score ≥ 80 且所有维度 passed)
score: number; // 综合质量分 0-100(安全性维度不参与计算)
dimensions: {
design_token_compliance: DimensionResult;
structural_integrity: DimensionResult;
typographic_quality: DimensionResult;
genre_fit: DimensionResult;
decoration_usage: DimensionResult;
security: DimensionResult; // 一票否决维度
};
actionable_feedback: string[]; // 可操作的修正建议列表(优先级排序)
}
interface DimensionResult {
passed: boolean;
score: number; // 该维度得分 0-100
issues: string[]; // 发现的问题列表
}
What ships with it
6 files 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 First seen · 131 lines · 165 tokens per session scan A 7eba67dbde71
html-review is a skill published in the GitHub repository rojim666/SztuCode (64 stars, last pushed today), licensed MIT. It adds 165 tokens to every session and 1,698 once invoked, about $0.0008 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 2026-09-12.
Other skills, from other repositories
skill-test
Verifies that the skill system is working correctly.
verify
Drive the MCP extension through the real Pi TUI and capture provider-prefix behavior.
visual-acceptance
A visual final-review method for user-interface changes. It checks the rendered result with repeatable screenshots, pixel values, computed styles, and CSS layering checks across themes.
design-qa-checklist
Build a QA checklist for verifying that a build matches the design. Use at implementation review. For the spec engineers build from, use handoff-spec.
ui-testing
Visual testing - catch invisible buttons, broken layouts, contrast.
visual-validation
Autonomous screenshot verification — captures, compares, and AI-evaluates UI changes for visual regressions.