dym-diagram-design
Warn
Audited by Gen Agent Trust Hub on Oct 17, 2026
Risk Level: MEDIUMDYNAMIC_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [DYNAMIC_EXECUTION]: The skill operates as a hub that dynamically resolves and loads sub-skill files based on a user-supplied
<slug>argument. It constructs file paths such as../dym-diagram-design-<slug>/SKILL.mdand.overstack/doyourmagic/diagram-design/skills/dym-diagram-design-<slug>/SKILL.mdto load instructions. Dynamic loading from computed paths based on external input is a known vulnerability that can allow for unauthorized file access or execution of unintended instructions. - [INDIRECT_PROMPT_INJECTION]: The skill creates an attack surface for indirect prompt injection by using unvalidated user input to determine the execution context.
- Ingestion points: The user-provided
<slug>parameter in the/dym-diagram-designcommand. - Boundary markers: Absent. The skill does not explicitly instruct the agent to verify that the slug matches the predefined list in the documentation before attempting to read the file.
- Capability inventory: The hub directs the agent to follow all instructions ("Làm theo Steps/Rules") within the resolved file, effectively delegating its full processing capability to the content of the target file.
- Sanitization: Absent. There are no instructions to sanitize the input slug for path traversal characters (e.g.,
../) or to ensure the resulting path remains within an expected directory structure.
Audit Metadata