guidewire-common-errors
Pass
Audited by Gen Agent Trust Hub on Mar 12, 2026
Risk Level: SAFEPROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill provides links to official Guidewire documentation and Gosu language resources (docs.guidewire.com, gosu-lang.github.io) to assist users with further troubleshooting.
- [PROMPT_INJECTION]: The skill is designed to process and analyze external error logs and API responses provided by the user. This creates an attack surface for indirect prompt injection if the data being analyzed contains instructions intended to override agent behavior.
- Ingestion points: External error logs and status code responses processed during diagnosis (SKILL.md).
- Boundary markers: No specific delimiters or instructions are provided to the agent to ignore embedded commands within the logs.
- Capability inventory: The skill has access to Read, Write, Edit, Grep, and Bash tools (SKILL.md).
- Sanitization: There is no evidence of log sanitization or validation before processing.
Audit Metadata