write-script-java
Pass
Audited by Gen Agent Trust Hub on Oct 14, 2026
Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill instructs the agent to utilize CLI tools such as
wmill script previewto execute Java code for testing purposes andwmill generate-metadatato synchronize local project configurations. These are standard operations for the Windmill developer environment. - [EXTERNAL_DOWNLOADS]: The documentation explains how to declare Maven dependencies using specific comment syntax (
//requirements:) within Java files. These dependencies are resolved and downloaded by the platform during script execution or metadata generation. - [INDIRECT_PROMPT_INJECTION]: The instructions describe the implementation of preprocessor scripts that ingest untrusted data from various external sources including webhooks, emails, and message queues.
- Ingestion points: Preprocessor scripts in
SKILL.mdare designed to receive raw trigger data from external sources like webhooks, SQS, and emails. - Boundary markers: No specific instructions are provided to the agent regarding the use of delimiters or boundary markers when processing this external data.
- Capability inventory: The skill environment supports local file system updates via
wmill generate-metadataand script execution viawmill script preview. - Sanitization: The guidelines do not explicitly mandate sanitization techniques for the external input processed by these scripts.
Audit Metadata