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/xbuilderlab/cheat-on-content/cheat-learn-fromnpx skills add XBuilderLAB/cheat-on-content --skill cheat-learn-fromgit clone --depth 1 https://github.com/XBuilderLAB/cheat-on-contentWhat 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 | $0.00000 | $0.04532 |
| Opus 5 | $0.00000 | $0.02266 |
| Sonnet 5 | $0.00000 | $0.00906 |
| Haiku 4.5 | $0.00000 | $0.00453 |
Grade A, and why
cheat-learn-from 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 3d 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.
Copies of this mod
1 near-identical copy found in the catalogue:
- cheat-learn-from — 100% identical, 0 lines differ
How it starts
The opening of the file, as written. The whole thing — 400 lines — stays where its author put it; the contents beside it link to each section on GitHub.
/cheat-learn-from — 对标账号导入
工具早期最重要的信号源是对标账号——你 init 完没数据,rubric 等权 v0 等于占星。但如果你能找一个你想做成那样的账号,导入 5-10 条它的高/中/低样本,工具就有了 anchor。
后期当你自己 calibration_samples ≥ 10 时,benchmark 影响自然减弱——你的真实数据成为主要信号源。但 benchmark.md 不删,仍是 cheat-seed brainstorm 的 reference frame。
Overview
[用户:学这个账号 / 启动 cheat-learn-from]
↓
[Phase 0: 检查 benchmark 状态]
↓
[Phase 1: 选 input 方式(Way a 默认)]
↓
[Phase 2: 收集材料]
Way a: 用户粘 N 条 script 文本 + 数据
Way b: 用 whisper 转录 samples/ 目录里的视频
↓
[Phase 3: 询问每条样本的"印象判断"(高/中/低 + 为啥)]
↓
[Phase 4: Claude 拆 pattern + 派生 rubric 信号]
↓
[Phase 5: 用户 review → 改 → 落盘]
↓
[Phase 6: 写 benchmark.md / script_patterns.md / rubric_notes.md]
↓
[Phase 7: 更新 state.benchmark_status]
Constants
- MIN_SAMPLES = 3 — 最少 3 条样本(少于拆不出 pattern)
- RECOMMENDED_SAMPLES = 5-10 — 推荐区间,平衡信号量 vs 用户工作量
- MAX_SAMPLES_PER_RUN = 15 — 单次导入上限——再多 Claude context 不够 + 用户也累
- DEFAULT_WAY = a — Way a 简单 + 准确,是 default
Inputs
| 必填 | 来源 |
|---|---|
<账号名> |
用户参数;缺失则询问 |
.cheat-state.json |
状态文件 |
| Way a: 用户粘的 script 文本 + 数据 | 对话 |
Way b: samples/<账号名>/*.mp4 等视频文件 |
用户提前下载好放进去 |
Workflow
Phase 0: 检查 benchmark 状态
读 .cheat-state.json 的 benchmark_status:
| 状态 | 处理 |
|---|---|
none |
首次导入——继续 Phase 1 |
pending |
用户之前答应等下找——继续 Phase 1 |
imported 已有 benchmark |
询问"你已有 benchmark [当前名],N 条样本。要做什么? a) 追加新视频到当前 benchmark b) 替换为新 benchmark c) 只看不改" |
参数解析:
--append→ 追加到现有 benchmark--replace <new-name>→ 用新 benchmark 替换(旧的归档到 benchmark.archived/)- 没标志 + 已有 benchmark → 走上面询问
Phase 1: 选 input 方式(两个独立维度)
每条样本 = script + 数据。两者怎么拿是独立的——你可以混搭。
Phase 1a: script source(怎么拿稿子)
script 怎么拿?
a) **粘文本(最简单,推荐)**
- 你自己整理过 / 用工具提取过——直接粘进对话
- 工具推荐(按方便程度排):
抖音 / 小红书:
- 微信小程序「轻抖」—— 粘视频链接 → 自动提取文案 + 评论。最快
- 类似工具:"视频解析助手" / "短视频文案提取" 等小程序
- 通常有免费额度,重度使用收费
B 站 / YouTube:
- 视频页面有"显示字幕/文字记录"按钮(如果 UP 主开了)
- 第三方:DownSub / SaveSubs / yt-dlp --write-auto-sub
公众号 / Substack:
- 直接复制网页文字
b) **whisper 转录视频文件**
- 你下载了视频到 samples/<账号名>/<video>/source.mp4
- 需要装 whisper-cpp + ffmpeg(见 adapters/script-extraction/whisper/README.md)
- 转录可能有错别字 / 漏字 / 标点不准——准确度比 a 差
c) **跳过 script,只用元数据 + 印象**
- 你拿不到稿子也懒得用工具
- 后果:pattern 拆不深(只能看标题 / 数据 / 你的印象),但 rubric 信号还行
- 适合"先快速搭起来,将来补"
回 a / b / c。
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.
- 3d ago First seen · 400 lines · 0 tokens per session scan A ca9aa5d7b8f1
cheat-learn-from is a skill published in the GitHub repository XBuilderLAB/cheat-on-content (6,715 stars, last pushed 3d ago), licensed MIT. It costs nothing until one of its globs matches a file; then it loads 4,532 tokens. 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
systematic-debugging
Use when encountering any bug, test failure, or unexpected behavior, before proposing fixes.
brainstorming
You MUST use this before any creative work - creating features, building components, adding functionality, or modifying behavior. Explores user intent, requirements and design before implementation.
auto-perf-optimize
Run agent-driven VS Code performance or memory investigations. Use when asked to launch Code OSS, automate a VS Code scenario, run the Chat memory smoke runner, capture renderer heap snapshots, take workflow screenshots, compare run summaries, or drive a repeatable scenario before heap-snapshot analysis.
chat-perf
Run chat perf benchmarks and memory leak checks against the local dev build or any published VS Code version. Use when investigating chat rendering regressions, validating perf-sensitive changes to chat UI, or checking for memory leaks in the chat response pipeline.
chat-pet-sprite-creation
Use when creating or changing VS Code chat pet sprite art, sprite sheets, state animations, eye treatments, Stable/Insiders variants, or pet transitions under src/vs/workbench/contrib/chat/browser/widget/media/chatPet.
cpu-profile-analysis
Analyze V8/Chrome CPU profiles (.cpuprofile) and DevTools trace files (Trace-.json). Use when: profiling performance, investigating slow functions, comparing code paths, finding bottlenecks, analyzing timeToRequest, understanding call trees from sampling profiler data, analyzing layout/paint/rendering, investigating…