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 commands/xiaobei930/cc-best/traingit clone --depth 1 https://github.com/xiaobei930/cc-bestWhat 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 | $0.00008 | $0.00589 |
| Opus 5 | $0.00004 | $0.00295 |
| Sonnet 5 | $0.00002 | $0.00118 |
| Haiku 4.5 | $0.00001 | $0.00059 |
Grade A, and why
train 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 2d 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.
What it actually says
/train - 训练模型
训练项目中的机器学习/深度学习模型。
适用场景
- 模型微调
- 迁移学习
- 从头训练
通用训练流程
1. 环境检查
# GPU 检查
python -c "import torch; print(f'CUDA: {torch.cuda.is_available()}')"
nvidia-smi
# 依赖检查
pip list | grep -E "torch|tensorflow|transformers"
2. 数据准备
- 检查训练数据集是否就绪
- 验证数据格式和质量
- 划分训练/验证/测试集
3. 训练配置
# 示例配置结构
model:
name: "model_name"
pretrained: true
training:
epochs: 10
batch_size: 32
learning_rate: 1e-4
data:
train_path: "data/train"
val_path: "data/val"
4. 启动训练
# 通用启动命令
python train.py --config config/train.yaml
# 使用 GPU
CUDA_VISIBLE_DEVICES=0 python train.py
# 后台运行
nohup python train.py > train.log 2>&1 &
训练监控
日志记录
- Loss 变化曲线
- 验证集指标
- 学习率变化
Checkpoint 管理
- 定期保存模型权重
- 保留最优模型
- 支持断点续训
常用框架
| 框架 | 训练命令 |
|---|---|
| PyTorch | python train.py |
| Hugging Face | python -m transformers.trainer |
| TensorFlow | python train.py |
| PyTorch Lightning | python train.py |
注意事项
- 确保有足够的 GPU 显存
- 使用 tmux/screen 后台运行长时间任务
- 定期保存训练状态和 checkpoint
- 记录实验参数和结果
项目定制
根据项目需要修改此文件，添加具体的：
- 模型类型和参数
- 数据集路径
- 训练脚本位置
- 评估指标
记住: 训练是投资而非成本——花 10 分钟训练 AI 理解项目规范，省下数小时的手动纠正。
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.
- 2d ago First seen · 106 lines · 8 tokens per session scan A 30ccab3a554f
train is a command published in the GitHub repository xiaobei930/cc-best (50 stars, last pushed 2mo ago), licensed MIT. It adds 8 tokens to every session and 589 once invoked, about $0.0000 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 commands, from other repositories
al
Run AgentLint diagnostic across all projects. Use when: user says /al, 'check all projects', 'agent lint', or '体检'.
tldr
Re-apply TLDR rules for this turn (verdict first, no filler).
moyu-lite
Invoke the moyu:moyu-lite skill and follow it exactly.
audit-plugin
Audit plugin skills, commands, and agents for structure, size, and naming issues.
lfe-dep-audit
Inspector sub-skill. Reviews dependency manifest files (package.json, requirements.txt, go.mod, Cargo.toml, pom.xml) changed in the current diff for risky version patterns and stale majors. Emits a human-run audit instruction rather than executing tools. Writes .plans/checks/depfindings.md. Called by lfe-inspector…
lfe-plan-critique
Run a 5-lens pre-build critique of the approved active plan before the Builder starts. Acts as the Architect persona, read-only on src/. Writes .plans/plancritique.md. Use immediately after Brain approves activeplan.md.