webflow-cli:troubleshooter

Pass

Audited by Gen Agent Trust Hub on Mar 11, 2026

Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill executes local system commands including webflow CLI operations and filesystem modifications like rm -rf node_modules to resolve environment issues.
  • [EXTERNAL_DOWNLOADS]: The skill facilitates the installation and update of the @webflow/cli package from the npm registry. These are vendor-provided resources and part of the intended troubleshooting workflow.
  • [PROMPT_INJECTION]: The skill ingests untrusted error messages and logs provided by the user to determine the root cause of issues. This represents a surface for indirect prompt injection. 1. Ingestion points: User descriptions and captured error messages in SKILL.md. 2. Boundary markers: Absent. 3. Capability inventory: Arbitrary command execution via Bash tool and file reading via Read tool. 4. Sanitization: Absent; the agent relies on raw error output to choose repair actions.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 11, 2026, 06:34 PM
Security Audit — agent-trust-hub — webflow-cli:troubleshooter