webflow-mcp:custom-code-management

Pass

Audited by Gen Agent Trust Hub on May 10, 2026

Risk Level: SAFEPROMPT_INJECTION
Full Analysis
  • [PROMPT_INJECTION]: The skill demonstrates a surface for indirect prompt injection because it ingests and processes data from an external source (Webflow site scripts) which is then used to inform the agent's analysis and output.
  • Ingestion points: The skill uses list_registered_scripts, list_applied_scripts, and get_page_script to retrieve script names, source code, and versions from the Webflow environment.
  • Boundary markers: The instructions do not specify any delimiters or safety warnings for the agent to ignore potentially malicious instructions embedded within the retrieved script content.
  • Capability inventory: The skill has the capability to modify site configuration through tools like add_inline_site_script, upsert_page_script, and delete_all_site_scripts.
  • Sanitization: There is no evidence of sanitization or filtering of the script content before it is processed by the agent during the Analysis and Verification phases.
Audit Metadata
Risk Level
SAFE
Analyzed
May 10, 2026, 01:51 AM
Security Audit — agent-trust-hub — webflow-mcp:custom-code-management