crowdin-context-cli
Pass
Audited by Gen Agent Trust Hub on Sep 16, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill defines a workflow where an AI agent downloads and processes external strings from Crowdin. These strings represent untrusted data that could contain hidden instructions designed to influence the agent's behavior during the 'ai_context' enrichment phase.
- Ingestion points: Data is retrieved from the Crowdin platform into a local
crowdin-context.jsonlfile via thecontext downloadcommand. - Boundary markers: While the JSONL format provides structural boundaries, the natural language content within the
textandcontextfields is interpreted by the agent without specific delimiters or warnings to ignore embedded instructions. - Capability inventory: The skill enables shell command execution through the
crowdinCLI for file system operations and network communication with the Crowdin API. - Sanitization: No sanitization or validation logic is specified for the content of the downloaded strings before the agent processes them.
- [COMMAND_EXECUTION]: The skill instructions involve executing the
crowdinCLI tool with various parameters, including sensitive configuration options like project IDs and API tokens, to manage remote project data.
Audit Metadata