webflow-local-dev-loop
Pass
Audited by Gen Agent Trust Hub on Sep 10, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to process external project data, which introduces a potential attack surface. However, it implements robust safety protocols to mitigate these risks:
- Ingestion points: The agent is instructed to read local project files such as
webflow.json, package manifests, and script files (SKILL.md). - Boundary markers: Instructions explicitly separate the discovery phase (read-only) from the mutation phase (write/edit), ensuring the agent does not act prematurely on discovered data.
- Capability inventory: The skill utilizes
WriteandEdittools to modify the project based on its findings (SKILL.md). - Sanitization: The instructions include mandatory redaction of sensitive data, forbidding the agent from echoing, persisting, or logging bearer tokens and other credentials (SKILL.md). It also requires explicit user approval for any destructive or production-facing operations.
Audit Metadata