omni-usage-logs

Pass

Audited by Gen Agent Trust Hub on Aug 21, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: The skill documentation describes interactions with a local API on localhost. It correctly uses shell environment variables ($OMNIROUTE_TOKEN) as placeholders for authentication, avoiding hardcoded secrets.
  • [INDIRECT_PROMPT_INJECTION]: The skill has a surface for indirect prompt injection as it retrieves call and request logs which may contain data from previous AI interactions.
  • Ingestion points: SKILL.md documents endpoints like /api/usage/call-logs that ingest log data.
  • Boundary markers: There are no explicit instructions to the agent to treat log content as untrusted data or use delimiters.
  • Capability inventory: The skill documentation only covers HTTP read operations; it does not define scripts with command execution or file-write capabilities.
  • Sanitization: No sanitization of ingested log data is mentioned in the documentation.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 21, 2026, 01:07 AM
Security Audit — agent-trust-hub — omni-usage-logs