common-observability

Pass

Audited by Gen Agent Trust Hub on Sep 5, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill is designed to analyze and instrument untrusted source code across multiple languages, creating a vulnerability surface where malicious instructions embedded in those files could influence agent behavior.
  • Ingestion points: The skill triggers on a wide variety of backend files including .ts, .go, .java, .kt, and .py files as defined in SKILL.md.
  • Boundary markers: There are no instructions for the agent to use delimiters or isolation techniques (like wrapping file content in XML tags) when processing these external files.
  • Capability inventory: The agent is instructed to perform complex modifications to service logic, middleware, and interceptors.
  • Sanitization: No validation or sanitization steps are defined for the code being ingested and processed.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 5, 2026, 05:45 PM
Security Audit — agent-trust-hub — common-observability