zalando-research
Pass
Audited by Gen Agent Trust Hub on Sep 21, 2026
Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill utilizes a local bash script (
scripts/crawlora.sh) to perform API requests. This script is well-hardened, featuring validation for allowed paths and methods, and prevents query parameter injection by using curl's URL-encoding features. It also securely manages the API key by writing it to a temporary, restricted configuration file that is automatically deleted on exit. - [EXTERNAL_DOWNLOADS]: The skill retrieves product, search, and category data from the vendor-controlled domain
api.crawlora.net. These network operations are intrinsic to the skill's function as a fashion marketplace research tool. - [INDIRECT_PROMPT_INJECTION]: The skill ingests data from external Zalando storefronts through the Crawlora API.
- Ingestion points: Data enters the agent's context through the output of
scripts/crawlora.shas documented inSKILL.md. - Boundary markers: The skill does not currently use specific boundary markers or instructions to isolate the JSON results from the API.
- Capability inventory: The skill has access to the local shell for running the
scripts/crawlora.shscript. - Sanitization: The helper script sanitizes all query arguments and rejects suspicious patterns like the '@' character which could be used to trigger unexpected file reads in curl.
Audit Metadata