skills/mukul975/anthropic-cybersecurity-skills/performing-adversary-in-the-middle-phishing-detection/Gen Agent Trust Hub
performing-adversary-in-the-middle-phishing-detection
Pass
Audited by Gen Agent Trust Hub on Apr 6, 2026
Risk Level: SAFEPROMPT_INJECTION
Full Analysis
- [SAFE]: The skill's Python scripts ('scripts/agent.py' and 'scripts/process.py') are designed for local log analysis. They use standard libraries to process JSON data and do not perform any network operations, file modifications, or credential extraction.
- [PROMPT_INJECTION]: The skill processes external sign-in logs and inbox rules which could theoretically contain malicious strings. This presents a potential surface for indirect prompt injection, but the scripts themselves only perform passive mathematical calculations and pattern matching, which is consistent with their defensive purpose.
- [SAFE]: There is a minor documentation inconsistency between the author names in 'SKILL.md' and the 'LICENSE' file, which does not pose a security risk.
Audit Metadata