skills/mukul975/anthropic-cybersecurity-skills/performing-user-behavior-analytics/Gen Agent Trust Hub
performing-user-behavior-analytics
Pass
Audited by Gen Agent Trust Hub on Apr 7, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The script
scripts/agent.pydisables SSL/TLS certificate verification (verify_certs: False) when connecting to Elasticsearch. This is a common practice in development or lab environments but is a best-practice violation for production systems as it could allow for person-in-the-middle attacks.\n- [SAFE]: The skill processes untrusted log data, creating an indirect prompt injection surface area. Evidence chain:\n - Ingestion points:
scripts/agent.pyqueries Elasticsearch;SKILL.mdprovides Splunk queries for various authentication and file access logs.\n - Boundary markers: None identified in the report output or scripts.\n
- Capability inventory: The script performs network requests to Elasticsearch and prints results to stdout; no dangerous system-level capabilities are invoked.\n
- Sanitization: No explicit sanitization of log data (e.g., user names) is performed before display.\n- [SAFE]: There is a minor metadata mismatch where the
SKILL.mdfile lists 'mahipal' as the author, while theLICENSEfile and vendor context attribute the skill to 'mukul975'. The license also contains a future-dated copyright (2026).
Audit Metadata