editing
Pass
Audited by Gen Agent Trust Hub on Oct 14, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONDYNAMIC_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill acts as a proxy, passing user-provided arguments directly to various downstream specialized skills without sanitization or boundary markers. This represents a potential surface where malicious instructions in user data could influence the behavior of the destination skill.
- Ingestion points: User-supplied words passed as arguments to sub-skills (SKILL.md).
- Boundary markers: None present; input is passed through directly.
- Capability inventory: Invokes other skills via tools or by reading local instruction files (SKILL.md).
- Sanitization: No validation or escaping is applied to the user-supplied arguments before they are passed through.
- [DYNAMIC_EXECUTION]: The skill dynamically resolves the path to neighboring skill instruction files and executes their contents based on user-driven routing logic.
- Pattern: The instruction to 'read //SKILL.md' and follow it exactly constitutes dynamic loading of instructions from computed local paths.
Audit Metadata