rails-hotwire-driver

Pass

Audited by Gen Agent Trust Hub on Jun 20, 2026

Risk Level: SAFECOMMAND_EXECUTIONDATA_EXFILTRATION
Full Analysis
  • [COMMAND_EXECUTION]: The skill executes several bash and Ruby scripts (such as req.sh, submit_form.rb, and flow.sh) to automate HTTP requests, form submissions, and cookie management. These operations are strictly limited to local hosts (e.g., localhost, 127.0.0.1, *.localhost) through explicit validation checks in the scripts.
  • [DATA_EXFILTRATION]: The skill includes functionality to read local Rails development logs (readlog.sh) for the purpose of extracting One-Time Passwords (OTP) and verification links. This behavior is documented as a development affordance and includes a security check that refuses to process any log file with 'production' in the file path to prevent accidental exposure of production secrets.
  • [EXTERNAL_DOWNLOADS]: The skill references the use of standard development tools and libraries, such as the nokogiri Ruby gem and the @playwright/mcp package from NPM. These are well-known resources in the Rails and testing ecosystems and are used for their intended purposes.
  • [PROMPT_INJECTION]: As the skill processes server-rendered HTML and log content, it theoretically presents a surface for indirect prompt injection. However, this risk is mitigated by the skill's focus on local development environments and its use of structured parsing rather than raw execution of server-provided data.
Audit Metadata
Risk Level
SAFE
Analyzed
Jun 20, 2026, 01:55 PM
Security Audit — agent-trust-hub — rails-hotwire-driver