makepad-platform
Pass
Audited by Gen Agent Trust Hub on Feb 17, 2026
Risk Level: SAFE
Full Analysis
- PROMPT_INJECTION (SAFE): The keyword 'CRITICAL' in the description metadata is used as a task prioritization indicator for the LLM rather than an attempt to bypass safety filters or ignore previous instructions.
- INDIRECT_PROMPT_INJECTION (SAFE): The skill instructions read from a local file ('references/platform-support.md'). While this creates an ingestion surface, the skill does not have any capabilities (network, shell, or file-write) to exploit this surface, and the data source is local and trusted.
Audit Metadata