web-design-guidelines
Warn
Audited by Gen Agent Trust Hub on Feb 16, 2026
Risk Level: MEDIUMEXTERNAL_DOWNLOADSPROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS] (LOW): The skill fetches instructions and rules from a remote URL.
- Evidence: Reference to
https://raw.githubusercontent.com/vercel-labs/web-interface-guidelines/main/command.mdinSKILL.md. - Trust Status: The
vercel-labsorganization is a trusted source, downgrading this specific finding to LOW. - [PROMPT_INJECTION] (MEDIUM): The skill implements an Indirect Prompt Injection surface by fetching operational instructions (rules and output formats) from an external URL.
- Ingestion Point:
SKILL.mdfetches content fromcommand.mdvia WebFetch to determine its behavior. - Boundary Markers: Absent. The agent is instructed to 'Apply all rules' and 'Output findings using the format specified' without validation or isolation delimiters.
- Capability Inventory: The skill possesses file system read access and network access.
- Sanitization: None. There is no validation of the content fetched from the remote source before the agent processes it as logic.
Audit Metadata