The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted external data from the materials/ directory, including pitch decks, audit reports, and kickoff transcripts (Step 1.2 in references/methodology.md). This data is used to inform the generation of the marketing plan. While this is necessary for the skill's purpose, it presents a surface for indirect prompt injection if malicious instructions are embedded in the analyzed documents. The skill does not explicitly mention sanitization or boundary markers for this untrusted content.
[DATA_EXPOSURE]: The skill handles sensitive marketing and financial data, including MRR, churn, and LTV retrieved via the Stripe MCP (references/ops-stack-mapping.md). It appropriately manages credentials by instructing the use of restricted keys and environment variables (e.g., AHREFS_API_KEY in .env) rather than hardcoding them.
[COMMAND_EXECUTION]: The skill recommends the use of CLI tools like agent-browser and defuddle for browser automation and research (references/ops-stack-mapping.md). These are legitimate utilities intended to support the skill's operational stack for a fractional CMO workflow.

marketing-plan