omni-context-rtk
Pass
Audited by Gen Agent Trust Hub on Oct 15, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill facilitates the processing of untrusted external data through features such as prompt compression testing and TOML schema importing. This creates a potential surface for indirect prompt injection attacks.
- Ingestion points: External prompt samples and TOML filter files.
- Boundary markers: None documented in the skill instructions.
- Capability inventory: Network interaction with local API endpoints via curl commands.
- Sanitization: No evidence of data sanitization or validation before processing external content.
Audit Metadata