Runs approved scripts in a sandboxed environment with SSH support and audit logging.
npx skills add https://github.com/corv89/shannot --skill shannot
Please help me install this Agent Skill. Skill: shannot Source: https://github.com/corv89/shannot/tree/main Command: npx skills add https://github.com/corv89/shannot --skill shannot
Run diagnostic scripts in a sandboxed environment to prevent unintended changes to the host, while enabling human oversight via MCP.
Run a diagnostic script in the sandbox and review the queued actions with shannot approve.
High-intent search queries and answers about installing and using this skill.
You can safely execute Python diagnostic scripts by running them inside a PyPy sandbox that intercepts syscalls to isolate host system operations, preventing unintended changes to your local environment.
Yes, you can run sandboxed diagnostic scripts on remote servers via SSH, as the sandbox supports both local and remote execution to isolate operations across different machines.
Yes, the sandbox integrates with Claude Desktop and Claude Code through MCP, enabling AI-assisted workflows where you can review queued actions and safely approve operations.
The interactive approval workflow provides per-operation visibility for sandboxed script execution, allowing you to review queued actions, configure profiles, and maintain audit logs of approved operations.
The PyPy sandbox isolates host operations through syscall interception, meaning it is primarily designed for diagnostic scripts and may restrict direct system access required by certain Python libraries or operations.
lordmacuExecute shell commands and transfer files over SSH on allowlisted hosts.
insentekValidate AI-executed code changes against proposals and architectural constraints.
redpanda-dataManage the Redpanda AI CLI plugin lifecycle and MCP integration for Claude workflows.
AyomiscoImplement ChainVM interfaces and deploy custom virtual machines on Avalanche.
