brandkitfetch
Fail
Audited by Gen Agent Trust Hub on Sep 28, 2026
Risk Level: HIGHEXTERNAL_DOWNLOADSINDIRECT_PROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill instructs the agent to download brand data from the author's API at brand.replynodes.com. An automated scanner flagged the curl command as a remote code execution risk; however, this is a false positive as the command saves output to brand.json (a data file) and does not execute it. As the resource belongs to the skill author, it is considered standard functionality.
- [INDIRECT_PROMPT_INJECTION]: The skill retrieves brand information from external domains, which constitutes an untrusted data ingestion surface.
- Ingestion points: External brand data is fetched via the API described in SKILL.md.
- Boundary markers: The skill includes a defensive instruction to preserve the source domain and treat fetched content as data, not instructions.
- Capability inventory: The skill itself contains no executable scripts or system-level tools, though the agent's general capabilities may be influenced by the ingested data.
- Sanitization: There are no explicit instructions for the agent to sanitize or validate the content of the JSON response beyond general LLM safety guidelines.
Recommendations
- HIGH: Downloads and executes remote code from: https://brand.replynodes.com/stripe.com - DO NOT USE without thorough review
Audit Metadata