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 agentmods add skills/xiaotonng/pikiloom/snipenpx skills add xiaotonng/pikiloom --skill snipegit clone --depth 1 https://github.com/xiaotonng/pikiloomWrote 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/xiaotonng/pikiloom/snipe)<a href="https://agentmods.dev/skills/xiaotonng/pikiloom/snipe"><img src="https://agentmods.dev/badge/skills/xiaotonng/pikiloom/snipe.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 | $0.00079 | $0.02322 |
| Opus 5 | $0.00039 | $0.01161 |
| Sonnet 5 | $0.00016 | $0.00464 |
| Haiku 4.5 | $0.00008 | $0.00232 |
Grade A, and why
snipe 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 5d 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 — 141 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Snipe: Twitter/X 热帖截流
在 coding agent / AI 开发工具领域的高流量推广帖下回复 pikiloom，截取流量拿 star。
本 skill 只负责 Twitter 渠道的「发现 + 抽取 + 发布」机制。 所有产品话术、差异点、诚实边界、
语气/语言规则、回复骨架与反模式，统一来自 ../_promo/pitch.md —— 不在本文件
重复，也不要在本文件改话术（改话术只改 pitch.md）。去重 / 护栏 / 度量 / 发布姿态由共享核心负责：
| 关注点 | 位置 |
|---|---|
| 产品话术 SSOT | _promo/pitch.md（§9 表里 twitter 列是本渠道的机制差异） |
| 去重记录（写时强制） | _promo/registry.py（channel = twitter） |
| 发布护栏（频次/集中度/变体/熔断） | _promo/guard.py |
| 飞书推送 | _promo/push_feishu.py |
| 无人值守编排 + posture | _promo/orchestrate.md |
| 旋钮（caps / posture / kill switch） | _promo/config.json → channels.twitter |
运行根目录：
cd /Users/admin/Desktop/project/pikiloom。脚本前缀：.pikiloom/skills/_promo/。
Twitter 渠道的两个铁律（来自实测 + 平台研究）
- 链接绝不放主回复正文。 X 对带外链的帖/回复显著降权。主回复 = 一句差异化 +
npx pikiloom@latest（平台内原生文本，不触发降权）；GitHub 链接放到对自己回复的自回复（self-reply）里。 - 绝不复制同一段文案 / 同一条外链刷多帖 —— 这是 X 明文的封号信号。每条独立起草（
guard.py的 变体检查会拦截过相似的草稿）。
工作流（手动单次运行；无人值守见 _promo/orchestrate.md）
Step 1: 预检
cd /Users/admin/Desktop/project/pikiloom
python3 .pikiloom/skills/_promo/guard.py caps # 看 twitter 今日剩余配额；为 0 则今天不跑
Step 2: 搜索高流量推广帖（浏览器）
传入关键词参数 → 直接搜该词。否则用场景关键词动态发现（搜 3–5 组即可，不要死记产品名）：
coding agent tool / AI coding assistant / claude code tool / remote coding agent
AI agent dashboard / vibe coding 工具 / coding agent mobile / AI dev tool launch
claude code api cost / claude max subscription (billing 段 = 最高转化，命中用 pitch 角度 0)
对每个关键词：
browser_navigate→https://x.com/search?q={keyword}&src=typed_query&f=topbrowser_wait_for确认推文加载- 读取
scripts/extract_tweets.js，用browser_evaluate注入执行，得到 JSON - 结果不够就按 End 滚动一次再抽
- 合并所有关键词结果，按
text[:80]+url去重。无效关键词直接跳过。
Step 3: 筛选候选
硬门槛（全部满足）： views > 5000；最近 48h 内；与 pikiloom 有功能交集；正在推广具体产品/工具
（有 has_product_signal）；不在 registry：
python3 .pikiloom/skills/_promo/registry.py seen twitter "<tweet_url>" && echo SKIP # 已记录则跳过
排除： 自己(@sthnavy)的帖；纯新闻/讨论/提问；已是 pikiloom 用户；政治/争议。
What ships with it
3 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.
- 5d ago First seen · 141 lines · 79 tokens per session scan A 3d8009754b3b
snipe is a skill published in the GitHub repository xiaotonng/pikiloom (294 stars, last pushed 4d ago), licensed MIT. It adds 79 tokens to every session and 2,322 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 2026-08-30.
Other skills, from other repositories
performant-nodejs
Strategies for high-performance, scalable Node.js systems (Worker Threads, ESM, Observability, Clean Architecture).
debugging-output-and-previewing-html-using-ray
Use when user says "send to Ray," "show in Ray," "debug in Ray," "log to Ray," "display in Ray," or wants to visualize data, debug output, or show diagrams in the Ray desktop application.
review-pending-pr-reviews
Review, grill, edit, and post pending PR review drafts saved by /review-pr (or its batch/cron runners). Lists drafts that have not yet been posted, walks the maintainer through each finding, then posts, closes, or discards. Use when the user says "post my pending reviews", "what reviews are waiting", "go through the…
Cortex
Operate Cortex, the LifeOS memory system — the typed Knowledge Archive (People, Companies, Ideas, Research with typed related: links) plus recall of prior work sessions, ISAs, and conversations. Search, add, harvest, develop, ingest, distill, graph-navigate, recall. USE WHEN cortex, knowledge, knowledge base, search…
ApertureOscillation
3-pass scope oscillation that holds a question constant while shifting zoom — narrow/tactical, wide/strategic, then synthesis — to surface design tensions, scope recommendations, and coherence assessments invisible at any single zoom level. USE WHEN aperture oscillation, oscillate scope, zoom in and out, tactical vs…
RedTeam
Adversarial analysis deploying parallel expert agents to stress-test ideas, strategies, and plans — decomposes into atomic claims, attacks them, then steelmans and counter-argues, producing severity-ranked findings with remediation. USE WHEN red team, attack idea, counterarguments, critique, stress test, devil's…