skills/mukul975/anthropic-cybersecurity-skills/analyzing-ransomware-leak-site-intelligence/Gen Agent Trust Hub
analyzing-ransomware-leak-site-intelligence
Pass
Audited by Gen Agent Trust Hub on Sep 17, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill ingests and processes untrusted JSON data from third-party ransomware tracking services, which could be manipulated by threat actors to include malicious instructions targeting the AI agent.
- Ingestion points: Victim data is retrieved via
requests.get()from external sources includingransomwatch(GitHub) andransomware.livein bothSKILL.mdandscripts/agent.py. - Boundary markers: The skill lacks delimiters or specific instructions to ignore potential commands embedded within the external data strings.
- Capability inventory: The skill is capable of writing intelligence reports to local files (
ransomware_intel_report.md) and displaying processed data in the console. - Sanitization: No sanitization, escaping, or strict schema validation is applied to the retrieved content before it is interpolated into output reports.
Audit Metadata