skills/mukul975/anthropic-cybersecurity-skills/performing-cloud-forensics-investigation/Gen Agent Trust Hub
performing-cloud-forensics-investigation
Warn
Audited by Gen Agent Trust Hub on Sep 15, 2026
Risk Level: MEDIUMPRIVILEGE_ESCALATIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [PRIVILEGE_ESCALATION]: The skill contains instructions to execute commands with elevated privileges using
sudoto mount forensic evidence volumes. - Evidence:
sudo mount -o ro /dev/xvdf1 /mnt/evidenceinSKILL.md. This is a necessary operation for the intended forensic workflow but involves privileged execution. - [INDIRECT_PROMPT_INJECTION]: The skill ingests and processes external data from cloud logs (AWS CloudTrail, Azure Activity Log, GCP Audit Logs) which can contain attacker-controlled strings.
- Ingestion points: Forensic data is retrieved via
aws cloudtrail lookup-events,az monitor activity-log list,gcloud logging read, andboto3paginators inscripts/agent.py. - Boundary markers: The skill lacks explicit boundary markers or instructions to ignore embedded commands within the log data.
- Capability inventory: The skill has the capability to write to the file system, create cloud snapshots, and modify instance security group attributes.
- Sanitization: The Python scripts in
scripts/agent.pyandSKILL.mdparse JSON data but do not perform sanitization on string values (e.g., User-Agents, error messages) before including them in generated reports or displaying them.
Audit Metadata