The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill instructs the agent to read and incorporate information from a local context file, which creates a potential surface for indirect prompt injection if that file contains untrusted content.
Ingestion points: The agent is instructed to read .claude/product-marketing-context.md before asking initial assessment questions in SKILL.md.
Boundary markers: The instructions do not specify any delimiters or warnings to ignore embedded instructions within the context file.
Capability inventory: The skill is designed to generate analytics tracking plans, JavaScript implementation snippets, and strategic naming conventions.
Sanitization: No sanitization or validation logic is defined for the content ingested from the context file.
[EXTERNAL_DOWNLOADS]: The skill provides implementation templates that reference and load scripts from well-known services.
Evidence: The GTM implementation guide in references/gtm-implementation.md provides a template for the Facebook Pixel which fetches source code from https://connect.facebook.net/en_US/fbevents.js.

analytics-tracking