The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: Fetches components from the author's official GitHub repository and utilizes the standard npx skills installer from Vercel Labs. These are established and trusted methods for skill deployment.
[DATA_EXFILTRATION]: Accesses sensitive campaign data, including message templates, lead information, and historical conversation logs via the La Growth Machine MCP tools. This data is handled within the agent's context for the purpose of benchmarking and is not transmitted to unauthorized external domains.
[COMMAND_EXECUTION]: Implements interactive UI widgets that use the sendPrompt function to execute follow-up tasks. These tasks include campaign creation and sequence rewriting via the MCP tools, providing a streamlined workflow for the user.
[PROMPT_INJECTION]: Ingests and processes untrusted data from external message histories (e.g., sent emails or LinkedIn messages). This introduces a surface for indirect prompt injection. (1) Ingestion points: Fetches templates and conversation history via get_campaign_messages, get_lead_conversations, and get_conversation_messages. (2) Boundary markers: None identified in the prompt instructions to isolate external data from the analysis logic. (3) Capability inventory: Can trigger campaign creation (create_campaign) or prompt-based rewrites via sendPrompt in UI widgets. (4) Sanitization: No specific sanitization or filtering of external message content is mentioned before processing.

campaign-challenger