omni-context-rtk

Pass

Audited by Gen Agent Trust Hub on Aug 20, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: The skill defines interactions with a local API service on localhost:20128. All network operations are directed to the local loopback address, which is a whitelisted and safe destination.
  • [DATA_EXPOSURE_AND_EXFILTRATION]: Instructions use shell environment variables ($OMNIROUTE_TOKEN) for authentication tokens rather than hardcoding credentials, which is an appropriate security practice.
  • [INDIRECT_PROMPT_INJECTION]: The skill contains endpoints that ingest external data for processing, such as TOML schemas and text samples.
  • Ingestion points: Data payloads in POST /api/context/rtk/import and POST /api/context/rtk/test in SKILL.md.
  • Boundary markers: None identified in the skill instructions.
  • Capability inventory: HTTP interactions (curl) with local configuration and diagnostic endpoints.
  • Sanitization: Not explicitly documented; the skill assumes the underlying API handles validation.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 20, 2026, 09:52 PM
Security Audit — agent-trust-hub — omni-context-rtk