The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill is designed to read from local project files to gather context before providing analytics advice. This represents a surface where untrusted data (if project files are compromised) could influence the agent's instructions or implementation plans.
Ingestion points: SKILL.md instructions include reading .agents/marketing-strategy.md, .agents/marketing-learnings.md, and legacy strategy filenames.
Boundary markers: The skill does not define specific delimiters or instructions to ignore embedded commands within these context files.
Capability inventory: The skill generates tracking plans and implementation code for GA4 and GTM, including dataLayer pushes and custom JavaScript variables.
Sanitization: There is no mention of sanitizing or filtering content read from these external context files.
[EXTERNAL_DOWNLOADS]: The documentation includes implementation examples that reference remote scripts from well-known services.
Evidence: The GTM implementation guide provides the standard Facebook Pixel initialization script which loads https://connect.facebook.net/en_US/fbevents.js via a script tag.
Context: This is a standard, well-known analytics service and is documented as part of the primary purpose of the skill.

analytics