extension-to-functions-codebase
Pass
Audited by Gen Agent Trust Hub on Oct 5, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes external files, including
extension.yamland extension source code, to perform its migration tasks. This ingestion of untrusted data creates a surface for indirect prompt injection where malicious instructions could be embedded in metadata, descriptions, or code comments of the source files to influence the agent's output. - Ingestion points: Reads and parses
extension.yamland source files from the extension being migrated (SKILL.md). - Boundary markers: The instructions do not define explicit delimiters or instructions to ignore potential commands embedded within the processed data.
- Capability inventory: The skill generates new source code files,
package.jsonconfigurations, andREADME.mddocumentation. - Sanitization: There is no explicit requirement for the agent to sanitize or escape strings extracted from the extension's metadata before using them in generated code or documentation.
Audit Metadata