The Agent Skills Directory

[SAFE]: The skill is primarily instructional and configuration-based, defining workflows for the Intercom platform. It does not contain executable scripts or shell commands that pose a risk to the host environment.
[DATA_EXPOSURE]: The skill defines templates that use placeholders such as {{first_name}} and {{email}}. These are standard variables for personalization in customer support tools and do not involve unauthorized data access or exfiltration.
[INDIRECT_PROMPT_INJECTION]: As an automation tool for customer messaging, the skill naturally processes inbound user content.
Ingestion points: Inbound customer messages and conversation context are processed to determine bot replies or routing.
Boundary markers: None explicitly defined for variables like {{first_name}}, though these are typical template variables.
Capability inventory: The skill uses intercom_message and intercom_conversation tools to interact with users.
Sanitization: No explicit sanitization logic is described, but the risk is consistent with standard support automation use cases and is generally mitigated by the platform's internal safeguards.

Intercom Automation