The Agent Skills Directory

[SAFE]: The skill's instructions and associated Python scripts were audited for malicious patterns. No evidence of prompt injection, data exfiltration, obfuscation, or unauthorized access was found.
[COMMAND_EXECUTION]: The skill instructs the agent to execute local Python scripts (scripts/assess_health.py, scripts/fleet_survey.py, scripts/render_run_report.py) to format diagnostic data. These scripts were manually reviewed and found to be safe: they utilize only Python standard library modules (e.g., json, argparse, datetime), perform no network operations, and do not use dynamic execution sinks like eval() or exec().
[INDIRECT_PROMPT_INJECTION]: The skill acts as an observability interface for external agent logs and memory entries. While this constitutes an ingestion surface for untrusted data, the provided scripts are used exclusively for read-only terminal formatting and do not influence the agent's logic or security posture.

exploring-scouts