observe

Pass

Audited by Gen Agent Trust Hub on Apr 8, 2026

Risk Level: SAFECOMMAND_EXECUTIONCREDENTIALS_UNSAFEEXTERNAL_DOWNLOADSPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill uses python3 -c snippets and shell loops to manage query results, including batching trace IDs into temporary files and aggregating statistics. These operations are performed locally to facilitate complex data analysis.
  • [CREDENTIALS_UNSAFE]: Interacts with the ~/.config/observe.yaml file to read authentication tokens and tenant configuration. This is the standard method for the Observe CLI to maintain session state and authenticate requests.
  • [EXTERNAL_DOWNLOADS]: Provides instructions for installing the observe CLI tool from the official observeinc GitHub repository using the standard go install method.
  • [PROMPT_INJECTION]: Operates on observability data (logs, spans, etc.) which may contain untrusted content from external systems. This presents a surface for indirect prompt injection as the agent parses and interprets this data without explicit sanitization or boundary marker instructions.
Audit Metadata
Risk Level
SAFE
Analyzed
Apr 8, 2026, 02:05 AM
Security Audit — agent-trust-hub — observe