The Agent Skills Directory

[SAFE]: The skill serves as a set of operational guidelines for using a sandboxing tool called 'Sandcastle'. It promotes isolation by running tasks in separate branches and containers, which is a security-positive design for handling potentially risky changes or parallelizing development tasks.
[SAFE]: Instructions include mandatory cleanup procedures, requiring the agent to destroy sandboxes after use, even in the event of failure, which prevents resource exhaustion and stale environment persistence.
[SAFE]: The skill implements human-in-the-loop checkpoints for conflict resolution during the merge process, ensuring the user maintains control over code changes.
[SAFE]: No malicious patterns such as prompt injection, data exfiltration, or obfuscated code were detected. The skill relies on external configuration and setup (referenced as 'sandcastle.config.ts' and '/setup-sandcastle') rather than embedding executable logic directly.

sandcastle