correlate-observability-signals

Pass

Audited by Gen Agent Trust Hub on Mar 18, 2026

Risk Level: SAFEPROMPT_INJECTION
Full Analysis
  • [SAFE]: The skill provides legitimate technical documentation and configuration templates for metrics, logs, and traces. No evidence of malicious code, obfuscation, or unauthorized data exfiltration was detected.
  • [PROMPT_INJECTION]: Analysis of the indirect prompt injection surface shows the skill processes external observability signals. This is an inherent risk factor in the observability domain and no specific vulnerabilities were identified.
  • Ingestion points: Data enters the agent's context through logs from Loki, metrics from Prometheus, and traces from Tempo/Jaeger as described in SKILL.md.
  • Boundary markers: The templates do not define explicit isolation or "ignore instructions" markers for the agent's analysis of ingested telemetry.
  • Capability inventory: The skill allows the agent to use Bash, Grep, Read, Write, and Edit tools when interacting with these signals.
  • Sanitization: The provided instrumentation examples promote the use of structured logging and specific trace ID formats, which helps distinguish data fields from free-text content.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 18, 2026, 07:13 AM