marketing-retention

MEDIUM W011: Third-party content exposure detected (indirect prompt injection risk).

• Third-party content exposure detected (high risk: 1.00). The skill's Research Mode and the Research Playbook (SKILL.md and references/research-playbook.md) explicitly direct the agent to use agent-browser to fetch and read live public websites (e.g., churnkey.co, prosperstack.com, reallygoodemails.com, baremetrics.com, stripe.com) and to interpret that content to inform retention decisions, so it clearly ingests untrusted third‑party web content that can influence subsequent actions.

MEDIUM W012: Unverifiable external dependency detected (runtime URL that controls agent).

• Potentially malicious external URL detected (high risk: 0.90). The skill's runtime research/setup instructions tell the agent to install and run an external skill with "npx skills add https://github.com/vercel-labs/agent-browser --skill agent-browser" (and related npm install commands), which fetches and executes remote code that the skill relies on for live research, so this external GitHub URL is a runtime dependency that can directly affect agent behavior.

MEDIUM W009: Direct money access capability detected (payment gateways, crypto, banking).

• Direct money access detected (high risk: 1.00). The skill explicitly references and instructs using specific payment gateway and billing-platform features (Stripe, Stripe Billing Smart Retries, Stripe Portal, Card Updater/account updater, Braintree, Chargebee, Recurly, Churnkey integrations, etc.) and prescribes enabling/configuring them (smart retries, card updater, dunning sequences, retry schedules, updating payment methods). These are concrete payment-recovery and billing integrations — not generic browser automation or generic API callers — and therefore constitute direct financial execution capability.