The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted lead data from multiple external marketing channels to automate routing and outreach. This creates a surface where malicious instructions embedded in lead fields could potentially influence the agent's behavior.
Ingestion points: Untrusted lead data is captured from sources including events-webinars, paid-ads-manager, content-marketing, and community-builder (Phase 1.1).
Boundary markers: The skill does not define specific delimiters or instructions to treat lead-provided data as strictly data, which could allow for instruction confusion during prompt interpolation.
Capability inventory: The skill possesses the capability to trigger Slack/email notifications (Phase 3.1) and generate personalized email outreach (Phase 4.2), which are potential targets for data-driven manipulation.
Sanitization: There is no evidence of sanitization, filtering, or validation of the lead-supplied strings (such as source_detail or company) before they are used to generate personalized sentences in email templates.

lead-routing