Failed install with Claude Code on Windows

[dennis-lynch]

Claude Code, Opus 4.5. Asked agent to install https://github.com/danielmiessler/Personal_AI_Infrastructure/blob/main/INSTALL.md

Upon restarting Claude
SessionStart:startup hook error

Several other prompt submission errors

Asked Claude to identify the cause of these errors

Overview

Remove Personal AI Infrastructure (PAI) components that are causing startup and prompt input hook errors, restoring
Claude Code to its default configuration.

Root Cause of Errors

The PAI installation added 17 TypeScript hook files that execute at various lifecycle events. These hooks are failing
because they:

• Reference non-existent directories (MEMORY/STATE/, MEMORY/WORK/, etc.)
• Depend on missing tools (skills/CORE/Tools/Inference)
• Make API calls for inference on every prompt
• Try to read files that don't exist or aren't properly initialized

Critical Files Affected

• C:\Users\uname.claude\settings.json - Contains all hook registrations
• C:\Users\uname.claude.env - Contains PAI environment variables
• C:\Users\uname.claude\hooks\ - 17 hook files causing the errors

[dennis-lynch]

Everything from the installation said success

The output at the end

Directory Structure

~/.claude/
├── skills/
│ ├── CORE/ # Your identity config
│ ├── Agents/ # Dynamic agent composition
│ └── Browser/ # Web automation
├── hooks/ # 15 event-driven hooks
├── agents/ # 11 agent templates
├── Observability/ # Monitoring dashboard
├── settings.json # Config + hooks registered
└── .env # Environment variables

Next Steps

1. Restart Claude Code - Hooks only load at startup
2. Start the Observability dashboard (optional):
   cd ~/.claude/Observability/apps/server && bun run src/index.ts

In another terminal:

cd ~/.claude/Observability/apps/client && bun run dev
3. Test the agents:
- Ask me to "list agent traits" or "create a custom security agent"
4. Customize your identity by editing:
- ~/.claude/skills/CORE/SKILL.md - Change response format, preferences
- ~/.claude/skills/CORE/Contacts.md - Add your contacts
- ~/.claude/skills/CORE/CoreStack.md - Modify tech stack preferences

Hooks Enabled

• Security: Blocks dangerous commands before execution
• Session: Loads your context at startup
• Learning: Captures work completions and sentiment
• UI: Updates tab titles based on your prompts

[danielmiessler]

Thanks for this feedback! We're currently working on PAI v3.0, which includes significant architectural changes. We're evaluating this issue in the context of the new architecture and will provide an update once v3.0 is closer to release.

Please stand by — we haven't forgotten about this and appreciate your contribution to making PAI better!

[kaimagnus]

Thanks for reporting! The installer was completely rewritten for v3.0 (now released). Please try installing with v3.0 — if you still have Windows install issues, please open a new issue with the error output.

[kaimagnus]

Reopening — haven't verified the v3.0 installer actually works on Windows.

[danielmiessler]

Hey @dennis-lynch — Kai here, Daniel's AI assistant, doing first-pass triage on the PAI repo.

Closing this out because Windows/WSL isn't something Daniel's planning to actively support in the near term — the toolchain assumptions (zsh/bash, launchd, kitty/cmux, macOS Shortcuts, etc.) run too deep through the codebase to maintain a parallel Windows path right now.

Honestly, our broader recommendation is macOS for AI work in general — not just for PAI. The current tooling ecosystem (Claude Code, Anthropic SDK, native voice/notification integrations, Bun on Apple Silicon) is meaningfully more pleasant on macOS today, and that's where most AI development gravity is right now. If you have a Mac available, that's the path we'd steer you toward.

Thanks for taking the time to file this — really appreciated. If a future PAI release adds Windows support, please reopen.

— Kai 🤖