These reports document three experiments prompted by Andrew Simonet's question about whether writing could be protected from AI collection in a way analogous to Glaze or Nightshade for images.
| #Requires AutoHotkey v2.0 | |
| #SingleInstance Force | |
| #UseHook True | |
| #Warn | |
| Persistent | |
| SetWinDelay(0) | |
| ; Cycle full-screen Microsoft Dev Box / Cloud PC sessions from the local host. | |
| ; Targets launched remote session windows (msrdc.exe), not the Windows App launcher. |
| name | interview-packet |
|---|---|
| description | Prepare an interview packet from a candidate interview folder plus a role library file. Use when the user wants a candidate packet reviewed, a resume normalized from PDF to Markdown, and interviewer-ready outputs such as an interview plan, coding problem guide, and note form. Load the role from hr/role-library and stop if the role file does not exist. |
| license | Proprietary. LICENSE.txt has complete terms |
Use this skill to turn a candidate's interview materials into a lightweight interviewer packet.
This is the parent issue. Sub-issues are the component-level work items.
Build the first real end-to-end DBAgent scenario: detect a missing index causing high CPU, present the finding to the user, get approval, create the index, and then demonstrate autonomy progression by auto-resolving the same class of problem when the user sets an "always allow" instruction.
This scenario is the vehicle to build the first floor of every DBAgent component — real code, real data flow, real infrastructure. Each component is scoped to exactly what this scenario needs, with clear boundaries on what's deferred.
| name | multi-model-review |
|---|---|
| description | Reviews PRs or code changes using multiple AI models, synthesizes findings, and interactively applies fixes. Uses GPT 5.2, Gemini 3 Pro, and Opus 4.5 for diverse perspectives. |
Review PRs or code changes across multiple AI models to get diverse perspectives, then interactively work through findings with the user.
A hearty, paprika-forward beef stew inspired by Hungarian goulash, Indonesian Rendang, but with plenty of improvisation. This recipe is intentionally flexible—treat it like a framework you can riff on.
Note: I believe this is more of a Pörkölt, another Hungarian dish, but I did have Goulash in Budapest similar to this recipe.
Inspired by these:
All times UTC-3 (ART)
FOSS4G and the climate crisis: Let's get to work
Rob Emanuele, Microsoft
Sometimes you want to keep a node in an Azure Batch pool up and running as you're debugging tasks so that you don't have to pay the waiting time to spin up and down a node. You may want to push a new image to ACR with the same tag to run the job based on fixes based on the last run. However, Azure Batch nodes don't pull an updated image by default, and will run with the same image if a task is run again even if ACR has been updated.
The best solution I've found to speed things up is: