create-app
Pass
Audited by Gen Agent Trust Hub on Sep 9, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTIONEXTERNAL_DOWNLOADS
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill instructs the agent to fetch external Markdown documentation from the vendor's website to provide context for API capabilities, manifest fields, and CLI commands. It treats these external files as the authoritative source for code generation, creating a surface where modified remote content could influence the agent's output.
- Ingestion points: Documentation URLs listed in SKILL.md (e.g., modules.md, manifest.md, reference/cli-commands.md, llms-full.txt).
- Boundary markers: The instructions do not specify any delimiters or safety markers for processing the fetched content.
- Capability inventory: The skill executes shell commands (npx, pnpm), modifies local project files, and performs deployment actions (publish) to a live environment.
- Sanitization: No sanitization or content validation steps are described for the ingested documentation.
- [COMMAND_EXECUTION]: The instructions utilize shell commands for project management, including scaffolding with 'npx @crowdin/serverless-apps-cli@latest create', managing dependencies with 'pnpm install', and deploying code with 'npx crowdin-serverless-apps publish'. These commands are used to register applications and upload bundle files to the Crowdin platform.
- [EXTERNAL_DOWNLOADS]: The skill downloads official CLI tools and project templates from the npm registry and fetches documentation from the vendor's GitHub-hosted site (crowdin.github.io). These are official resources provided by the vendor for developing applications on their platform.
Audit Metadata