dify-dsl-generator
Pass
Audited by Gen Agent Trust Hub on Sep 14, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted user requirements to generate structured YAML configurations (DSL) for the Dify platform, including LLM system prompts and Python code blocks. This creates a surface for injection where malicious inputs could influence the generated logic.
- Ingestion points: User input via business requirement descriptions in
SKILL.mdandREADME.mdare interpolated into the final YAML output. - Capability inventory: The generated output includes functional Python3 code nodes and complex LLM instructions. While the agent itself does not execute these, the resulting artifact is designed for execution in the target Dify environment.
- Boundary markers: The skill does not provide instructions to escape or sanitize user requirements, making it possible for requirements to interfere with the intended structure of the generated DSL.
- Sanitization: There are no explicit validation steps defined to ensure the generated code or prompts are safe from malicious content before delivery to the user.
Audit Metadata