The Agent Skills Directory

[SAFE]: The skill documents an intricate workflow for sub-agent orchestration. It uses platform-specific tools like Agent and MCP tools (worker_commit, dispatch_import, dispatch_reap) which are expected in this environment.
[SAFE]: No hardcoded credentials, malicious obfuscation, or unauthorized remote code execution patterns were detected.
[SAFE]: The skill emphasizes security boundaries, such as Base-clean beats, BASE GUARD checks (verifying worktree isolation and lineage), and the use of a bwrap (Bubblewrap) jail for worker confinement to prevent unauthorized file system access.
[SAFE]: Instructions for cd operations and path resolution (e.g., ISS-031) are related to the specific environment's hook triggers and do not represent privilege escalation or persistence mechanisms.

dispatch-agent