📚 resourceActivePromisingcritical

sandboxed.sh

Self-hosted orchestrator for AI autonomous agents. Run Claude Code & Open Code in isolated linux workspaces. Manage your skills, configs and encrypted secrets with a git repo.

Stars

450

Skill Type

⚙️ Infrastructure Operations

Quality Score

107/200

License

Unknown

Forks

Last Updated

Jun 11, 2026

Discovered

Mar 25, 2026

Validation

Passed

github.com/Th0rgal/sandboxed.sh

Quality Breakdown

107/ 200

Content Signals

○Gotchas/Edge Cases+40

✓Progressive Disclosure+30

○Trigger Description+20

○Verification/Safety+20

○Code Examples+15

✓Composability+15

Repo Health

✓Recent Activity+15

✓Scripts/Automation+10

✓Real Usage (Issues)+10

✓Single Responsibility+10

✓Config/Persistence+10

✓Install Instructions+5

Multi-platform bonus: +5 pts if tool supports 2+ platforms. Score derived from 12 structural signals — not stars or popularity.

Trust & Verification

critical

Manual security review required. Use with extreme caution.

Active

Updated within the last 90 days. Actively maintained.

Unverified skill. Always review source code before installing any skill from an unknown author.

Risk Assessment

Self-hosted orchestrator explicitly designed to spawn and manage autonomous AI agents without human approval gates between missions
Provides isolated Linux workspace execution with systemd-nspawn containerization enabling arbitrary command execution
Multi-agent spawning capability - designed to run Claude Code, OpenCode, and Amp runtimes concurrently
SSH/VPN access enablement mentioned in vision ('Give an agent SSH access to your home GPU through a VPN'), allowing remote code execution
Persistent configuration management via .claude/ directory with CLAUDE.md and settings.json for agent behavior tuning
Git-backed skill library system allowing agents to load and execute arbitrary pre-defined code modules