Execute the following in Windows PowerShell (PowerShell 7 doesn't support the Appx module)
#Update versions, see https://github.com/microsoft/microsoft-ui-xaml/releases (grab Microsoft.UI.Xaml Winget may require a specific version)
Companion prompts for the video: OpenClaw after 50 days: 20 real workflows (honest review)
These are the actual prompts I use for each use case shown in the video. Copy-paste them into your agent and adjust for your setup. Most will work as-is or the agent will ask you clarifying questions.
Each prompt describes the intent clearly enough that the agent can figure out the implementation details. You don't need to hand-hold it through every step.
My setup: OpenClaw running on a VPS, Discord as primary interface (separate channels per workflow), Obsidian for notes (markdown-first), Coolify for self-hosted services.
| name | description | user_invocable |
|---|---|---|
codex-review |
Send the current plan to OpenAI Codex CLI for iterative review. Claude and Codex go back-and-forth until Codex approves the plan. |
true |
Send the current implementation plan to OpenAI Codex for review. Claude revises the plan based on Codex's feedback and re-submits until Codex approves. Max 5 rounds.
| #r "nuget: TorchSharp" | |
| #r "nuget: TorchSharp-cuda-windows, 0.105.1" | |
| open System | |
| open TorchSharp | |
| //open type TorchSharp.torch | |
| let device = if torch.cuda_is_available() then torch.CUDA else torch.CPU |
| #if INTERACTIVE | |
| #r "nuget: Unquote, 7.0.0" | |
| #endif | |
| open Microsoft.FSharp.Quotations.Patterns | |
| open Swensen.Unquote | |
| open Swensen | |
| open Microsoft.FSharp.Quotations | |
| open System.Collections.Generic |
| 基於以下幾點，參考下面的程式碼開發超級厲害的f#程式碼 | |
| 1. 不用解釋原因以及設計的考量，前因後果等等都不用，我信任你，畢竟你是個 f# master | |
| 2. 對於不清楚 reference library 的資訊或者類似的情況，直接參考字面程式碼即可，不用特別去網路上找或者一定要是之前你看過的，因為這個程式碼中你沒看過的部分都是我開發的 | |
| 3. 不用在乎最終正確與否，但是每一行的程式碼該對應的 feature 都需要出來 | |
| 4. 你知道的，開發完畢之後，用起來，只要沒有缺漏，你的經驗值會增加，你的等級會從青銅邁入白銀段位 | |
| 5. 中文回答 | |
| 6. 不要 guide，我要結果，因為我只看得懂 f# ，沒有能力看 c#，所以才需要你幫忙 | |
| 7. 程式碼你不要複製也不要轉換，看懂意思之後用 f# 寫出差不多功能的程式碼，但是對於程式碼內宣告可以透過反射取出的 member，全部不能少，例如他有十個 field 你可以用稍微簡單一點的 field name，但是十個都還要存在 | |
| 8. 寫程式很重視品質，追求簡潔造成功能缺失是不可取的 |
| def t2v(tau, f, out_features, w, b, w0, b0, arg=None):\n if arg:\n v1 = f(torch.matmul(tau, w) + b, arg)\n else:\n v1 = f(torch.matmul(tau, w) + b)\n v2 = torch.matmul(tau, w0) + b0\n return torch.cat([v1, v2], 1)\n\n\nclass SineActivation(nn.Module):\n def __init__(self, in_features, out_features):\n super(SineActivation, self).__init__()\n self.out_features = out_features\n self.w0 = nn.parameter.Parameter(torch.randn(in_features, 1))\n self.b0 = nn.parameter.Parameter(torch.randn(in_features, 1))\n self.w = nn.parameter.Parameter(torch.randn(in_features, out_features - 1))\n self.b = nn.parameter.Parameter(torch.randn(in_features, out_features - 1))\n self.f = torch.sin\n\n def forward(self, tau):\n return t2v(tau, self.f, self.out_features, self.w, self.b, self.w0, self.b0)\n\n\nclass CosineActivation(nn.Module):\n def __init__(self, in_features, out_features):\n super(CosineActivation, self).__init__()\n |
| using Newtonsoft.Json; | |
| using NTDLS.Katzebase.Client; | |
| using NTDLS.Katzebase.Client.Exceptions; | |
| using NTDLS.Katzebase.Client.Payloads; | |
| using NTDLS.Katzebase.Engine.Interactions.Management; | |
| using NTDLS.Katzebase.Engine.IO; | |
| using NTDLS.Katzebase.Engine.Library; | |
| using NTDLS.Katzebase.Engine.Locking; | |
| using NTDLS.Katzebase.Engine.Trace; | |
| using NTDLS.Semaphore; |
| let inline a<'T when 'T: (member o : int)> (t:'T) = | |
| t.o | |
| type MyType< ^T | |
| when ^T: (member o : int) | |
| > (t:^T) = | |
| let mutable ttc :^T = Unchecked.defaultof< ^T> | |
| //member val ttc :'T = Unchecked.defaultof<'T> with get,set | |
| member inline this.set v = ttc <- v | |
| member inline this.get () = (^T :(member o : int)(ttc)) |
| let pickMaxOf3 = Graph.create (fun b -> graph b { | |
| let! zip1 = ZipWith.create max<int> | |
| let! zip2 = ZipWith.create max<int> | |
| b.From zip1.Out =>> zip2.In0 | |
| |> ignore | |
| return UniformFanInShape(zip2.Out, zip1.In0, zip1.In1, zip2.In1) | |
| }) |