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 nmhjklnm/yaohe --skill promotegit clone --depth 1 https://github.com/nmhjklnm/yaoheWrote 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/nmhjklnm/yaohe/promote)<a href="https://agentmods.dev/skills/nmhjklnm/yaohe/promote"><img src="https://agentmods.dev/badge/skills/nmhjklnm/yaohe/promote.svg" alt="Measured on agentmods" 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 | $0.00086 | $0.01421 |
| Opus 5 | $0.00043 | $0.00711 |
| Sonnet 5 | $0.00017 | $0.00284 |
| Haiku 4.5 | $0.00009 | $0.00142 |
Grade A, and why
promote 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 7d ago.
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 — 69 lines — stays where its author put it; the contents beside it link to each section on GitHub.
promote — 开源项目一键合规宣传
核心原则:一键 ≠ 一键全发。渠道分三级自动化,越权发帖 = spam = 毁项目声誉。
流程
1. 项目画像(必做,一切文案的源头)
硬性前置闸门,任一不过就停下问用户,不进入后续步骤:
gh repo view <owner>/<repo>必须成功(仓库公开可达)。本地路径则检查git remote -v有 GitHub remote 且已 push——稿件里的每个链接都必须真实可达,绝不写入死链- LICENSE 文件真实存在才可声明协议;README 说 MIT 但没有 LICENSE 文件 → 提醒用户先补
然后读 README 全文、描述、topics、star 数、最近活跃度,提炼:一句话定位(它是什么+替代/优于什么)、目标人群、3 个差异化亮点、演示素材(截图/GIF/demo URL)。
README 薄的判定(缺任一项即问用户补齐,不要编):① 一句话说清它是什么 ② 安装/使用方法 ③ 至少一个具体使用场景或示例。
2. 渠道选择
读 channels/ 下所有渠道文件的 frontmatter,按项目画像过滤:
- 语言匹配:投 EN 渠道的条件 = 有英文 README(或 README 本身是英文);不满足时该渠道仍可列出但标注「需先补英文文档」,用户坚持投则明示风险后照办
- 门槛匹配(如渠道要求 star 数 / 入门级 / dev-tool 类目)
- 检查
PROMO-LOG.md(项目根目录):同一渠道 30 天内有任何记录(含 rejected)即跳过,冷却期从最新一条记录的日期起算
用 AskUserQuestion 列出推荐渠道(含自动化等级标注)让用户勾选;用户明确说"全投"则投所有合规匹配渠道。
3. 逐渠道生成定制稿
对每个选中渠道,读其 channels/<id>.md 全文,严格按其 format_spec 生成:
- 禁止一稿多投:每渠道的标题风格、篇幅、语气、结构都不同(阮一峰自荐 ≠ Show HN 标题 ≠ V2EX 分享创造帖)
- 文案原则:说人话、给事实(能做什么、和现有方案差在哪),禁营销腔/AI 味(震惊体、emoji 堆砌、buzzword)
- 去 AI 味终审(必做):装了 deslop 类 skill 就调用它过一遍;没装则至少执行硬规则:禁破折号(——)、禁「不是…而是…」否定对比句、弯引号换直角引号「」、删段末总结句、删互联网黑话
- 配图(必做):几乎所有渠道带图命中率都更高。优先级:① 项目已有截图/GIF 直接用 ② 真实运行录屏/截图(终端工具录一段 demo)③ 用可用的图像生成能力(imagegen/codex 出图/自绘 SVG 转 PNG)设计一张说明产品原理的宣传图。图片提交到项目仓库(如
docs/assets/)用 raw.githubusercontent.com URL 引用,别用外部图床。渠道规格标注图片必备的(如 ruanyf-weekly),无图不投 - 全部稿件先写入
promo-drafts/<channel-id>.md落盘;文件首行固定为Title: <标题>(即使该渠道标题走gh --title参数不进 body),复盘时才对得上号
4. 分级执行
按渠道 frontmatter 的 automatable 字段:
- full(issue/PR 型):先
gh auth status确认已登录(未登录 → 让用户gh auth login,停在这里)→ 展示最终稿 → 用户确认一次 →ghCLI 提交 → 记录返回 URL。失败就报错停,不换渠道兜底。 - assisted(网页表单型):给出最终稿 + 提交 URL + 逐字段填写指引;Chrome 扩展已连接时按
references/browser-assist.md流程代开页面、代填字段(登录和提交按钮永远归用户)。 - manual-only(社区发帖型):只交稿件 + 建议发帖时间(见渠道文件)+ 红线提醒。任何情况下不代发。
5. 记录与复盘
把本次投递写入目标项目根目录 PROMO-LOG.md,固定表格 schema(机器可读,下次运行先读它):
| channel | date | draft | url | status |
|---|---|---|---|---|
| ruanyf-weekly | 2026-07-02 | promo-drafts/ruanyf-weekly.md | https://github.com/ruanyf/weekly/issues/NNN | submitted |
What ships with it
22 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.
- channels/_template.md 761 B
- channels/appinn.md 830 B
- channels/awesome-claude-code.md 1.7 KB
- channels/awesome-list.md 1.7 KB
- channels/dev-newsletters.md 967 B
- channels/devto.md 1.1 KB
- channels/eleduck.md 660 B
- channels/gitee-gvp.md 859 B
- channels/githubdaily.md 1.1 KB
- channels/hellogithub.md 2.4 KB
- channels/indiehackers.md 632 B
- channels/jike.md 676 B
- channels/launch-boards.md 1.1 KB
- channels/lobsters.md 726 B
- channels/oschina.md 929 B
- channels/product-hunt.md 1.2 KB
- channels/reddit.md 1.3 KB
- channels/ruanyf-weekly.md 3.1 KB
- channels/show-hn.md 1.6 KB
- channels/sspai-matrix.md 1.0 KB
- channels/v2ex-create.md 1.0 KB
- references/browser-assist.md 1.3 KB
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.
- 7d ago First seen · 69 lines · 86 tokens per session scan A 60d80fb20366
promote is a skill published in the GitHub repository nmhjklnm/yaohe (8 stars, last pushed 2mo ago), licensed MIT. It adds 86 tokens to every session and 1,421 once invoked, about $0.0004 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
delegate
Let Claude Architect route a versioned implementation spec through the trusted MCP runtime, independently review the Candidate Artifact, record a decision, and integrate only accepted bytes. Use for implementation delegation, Producer selection, or commitment-boundary review.
codex
Run the Codex CLI directly in the user's checkout for code analysis, refactoring, or automated editing without Claude Architect's verified delegation lifecycle.
commit
A guided process for committing code changes to Git. It checks the working tree, follows a conventional commit format, links a task when available, reviews the changes, and asks for approval.
orient
Internal skill for commands. Render the shared orientation snapshot - active task, what's in flight, what's next - tracker-universal. Do not trigger on user conversation; only when a command needs the snapshot.
actor-fetch-personality
Internal skill for commands. Fetch and validate personality from URL or file path for Actor agent. Do not trigger on user conversation - only when commands need personality loading.
explain
Use when user invokes /explain with a file path, directory path, function/class name, or natural language concept. Also triggers on "explain this", "how does X work", "walk me through". Produces a structured, layered explanation of what the code does, how it connects, and where to start if you need to change it. Do…