vscode-ext-localization
Pass
Audited by Gen Agent Trust Hub on Oct 4, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to process external documentation and local repository files to implement localization. This functionality creates an attack surface where malicious instructions embedded in the processed files or documentation could influence the agent's behavior.
- Ingestion points: The skill reads local extension files (SKILL.md, Workflow step 1) and fetches external documentation or model inference results via vendor tools (
sandbase_discover,sandbase_run). - Boundary markers: The skill lacks explicit prompt delimiters or 'ignore instructions' warnings when interpolating content from these files into the agent context.
- Capability inventory: The skill is authorized to perform file modifications (SKILL.md, Workflow step 3) and execute external API calls via the
sandbase_runtool. - Sanitization: No explicit sanitization or filtering of the ingested content is defined in the instructions.
Audit Metadata