security-digest
Pass
Audited by Gen Agent Trust Hub on Apr 29, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill implements a structured approach to security monitoring by correlating data from multiple UniFi platforms using batch processing tools.
- [SAFE]: Analysis and severity logic are transparently defined in external reference documents, facilitating easy auditing of the agent's behavior.
- [SAFE]: While the skill processes untrusted event data from external servers, the risk of indirect prompt injection is minimal as the skill does not use dangerous capabilities such as arbitrary file writes, network requests to unknown domains, or code execution.
- Ingestion points: Event data from protect_batch, access_batch, and unifi_batch in SKILL.md.
- Boundary markers: Not explicitly used.
- Capability inventory: Restricted to security-related read tools and report generation.
- Sanitization: Not explicitly implemented.
- [SAFE]: No evidence of obfuscation, hardcoded credentials, or unauthorized command execution was found in the skill's instructions or logic.
Audit Metadata