Pre-built .safespace configuration templates for Claude, GPT, and local models. Import into Agent Safehouse and run agents securely in seconds.
$ safehouse import code-review.safespace
✓ Template loaded: Code Review v1.3.0
✓ Permissions verified: 4 allow, 4 deny
✓ Sandbox ready — run 'safehouse start' to beginEach template is a .safespace config file with granular permissions tuned for a specific workflow.
Sandbox for AI-assisted code review. Read-only access to project files with network isolation.
Isolated environment for data transformation. Write access to output directory only.
Controlled network access for web scraping. Allowlisted domains with write to scrape output only.
Safe file manipulation sandbox. Scoped write access with delete protection.
Outbound API sandbox with domain allowlisting. No filesystem write access.
Pick a use-case preset from the gallery.
Claude, GPT, or Ollama/local models.
One-click download of the config file.
Import into Agent Safehouse and start securely.
Free for early adopters. Premium template packs coming soon.
“Finally, I can let Claude review my code without worrying about it touching things it shouldn't. The code review template just works.”
“I was hand-rolling sandbox configs for months. These templates saved me hours and they're more secure than what I had.”
“The permission breakdowns are exactly what I want to see. Transparent, granular, and easy to audit.”