dt-obs-tracing
Pass
Audited by Gen Agent Trust Hub on Apr 3, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: No malicious patterns or security risks were identified. The skill consists of instructional markdown and DQL snippets for observability tasks.
- [DATA_EXFILTRATION]: The skill documents queries for accessing network-related metadata such as client IP addresses and resolved server IPs. These operations are limited to data retrieval within the Dynatrace Grail platform and are consistent with standard distributed tracing practices.
- [PROMPT_INJECTION]: The skill teaches the agent how to process and parse untrusted data from exception messages and log content, which presents a surface for indirect prompt injection.
- Ingestion points: Analysis of
span.events(exception messages) and log content strings viafetch spansandfetch logscommands. - Boundary markers: Absent in provided DQL examples.
- Capability inventory: Platform-restricted DQL queries for data retrieval and aggregation.
- Sanitization: No explicit sanitization or escaping logic is included for the parsed strings; the agent relies on the platform's query engine constraints.
Audit Metadata