planar-binding
Pass
Audited by Gen Agent Trust Hub on Oct 11, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONNO_CODE
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill defines a workflow for interacting with external APIs and agents, creating a surface for indirect instructions if an external entity provides malicious content.
- Ingestion points: External entity names and binding targets defined in user instructions (SKILL.md).
- Boundary markers: The procedure explicitly requires restating 'no-touch boundaries' and success conditions before taking action.
- Capability inventory: The skill file consists only of instructional text and contains no executable code or script-based tool calls.
- Sanitization: The skill relies on procedural scoping and explicit user confirmation rather than programmatic data sanitization.
- [NO_CODE]: The skill consists entirely of instructional markdown and metadata without any attached scripts or executable code.
Audit Metadata