The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data originating from external Email Service Providers (ESPs), such as campaign subject lines, body content, and flow names, which could be manipulated to include malicious instructions.
Ingestion points: Data enters the agent's context through several MCP tool calls, including mcp__cogny__klaviyo__list_campaigns, mcp__mailchimp__tool_get_report, and mcp__rule__tool_list_campaigns.
Boundary markers: The instructions lack specific requirements for using delimiters or boundary markers to isolate external data strings from the agent's instruction set.
Capability inventory: The skill is authorized to use the Bash and Write tools, and can perform stateful operations via mcp__cogny__write_context_node and mcp__cogny__create_finding.
Sanitization: There is no mention of sanitizing or escaping the retrieved campaign metadata before it is interpolated into the narrative generation prompts for the Slack, CEO, and Deck outputs.

email-report