troubleshoot-apache-pulsar

Pass

Audited by Gen Agent Trust Hub on Jul 10, 2026

Risk Level: SAFEPROMPT_INJECTION
Full Analysis
  • [PROMPT_INJECTION]: The skill relies on external data ingested from Netdata via MCP, which creates a surface for indirect prompt injection.
  • Ingestion points: The agent is instructed to process outputs from MCP functions such as query_metrics, find_anomalous_metrics, list_metrics, and list_raised_alerts (as seen in SKILL.md and multiple rule files).
  • Boundary markers: The instructions do not specify the use of delimiters or 'ignore embedded instructions' markers when handling the external metric data or alert descriptions.
  • Capability inventory: The skill is restricted to read-only monitoring operations and providing diagnostic recommendations. It does not include scripts or tools for file system writes, network exfiltration, or arbitrary command execution.
  • Sanitization: There is no evidence of sanitization or validation of metric labels, metadata, or alert content before the agent processes them.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 10, 2026, 11:23 AM
Security Audit — agent-trust-hub — troubleshoot-apache-pulsar