webflow-data-handling
Pass
Audited by Gen Agent Trust Hub on Sep 10, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill ingests data from local repository files and external Webflow documentation to inform its design and implementation tasks. While this creates a theoretical surface for instruction injection from external sources, the risk is mitigated by the use of official documentation and strict human approval requirements.
- Ingestion points: Repository files via Read/Glob/Grep tools and official documentation via WebFetch.
- Boundary markers: Not explicitly defined for delimiting agent instructions from data.
- Capability inventory: Write and Edit tools are available for modifying project files.
- Sanitization: No specific sanitization of documentation content is defined, but the workflow emphasizes manual verification and user approval.
- [DATA_EXPOSURE]: The skill includes robust security guidelines to prevent the accidental exposure of sensitive information. It explicitly prohibits the agent from reading, echoing, or persisting authentication tokens (Bearer, site, or OAuth tokens) in any output, including logs and reports. It also requires redacting sensitive fields from event payloads and debug materials.
- [COMMAND_EXECUTION]: The skill follows a principle of least privilege by defaulting to read-only inspection. It mandates that any state-changing operations, such as deletions or deployments, must be preceded by a clear explanation of the proposed change and require explicit user authorization.
Audit Metadata