The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill's installation process requires cloning a repository from a third-party GitHub user (Chaitanyaa0406/AI-Festival-Campaign-Generator.git). This source is not associated with the skill author or a trusted organization, posing a supply chain risk.
[COMMAND_EXECUTION]: The setup guide includes shell commands for cloning the external repository. Users should manually audit the contents of the repository before execution as it originates from an unverified source.
[INDIRECT_PROMPT_INJECTION]: The skill defines several n8n code nodes that interpolate external data (from webhooks) directly into prompts for the Gemini model and into HTML email templates. This creates an attack surface where malicious input could influence the AI's behavior or inject malicious content into outgoing communications.
Ingestion points: Webhook trigger fields such as festival_name, product_name, and brand_name in SKILL.md (Pattern 1).
Boundary markers: Absent. The code snippets construct prompts by directly embedding variables (e.g., const prompt = \Festival: ${festivalName}...``) without delimiters or instructions to ignore embedded commands.
Capability inventory: The workflow has the capability to send emails via Gmail API, upload files to Google Drive, and write data to Google Sheets.
Sanitization: The provided validateCampaignData function only checks for the presence of fields and basic email formatting, providing no protection against prompt injection or XSS in HTML emails.

ai-festival-campaign-generator