The Agent Skills Directory

[PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection due to its architecture of ingesting and processing untrusted data from multiple sources to provide context-aware simulations.
Ingestion points: The skill reads project-specific context from local files such as .agents/product-marketing.md, .claude/product-marketing.md, and product-marketing-context.md. It also allows for the creation and loading of custom advisor dossiers and fetches external data via the "Live Research Pass" using web searches.
Boundary markers: There are no instructions to wrap ingested data in unique delimiters or to provide the agent with explicit instructions to ignore potentially malicious embedded commands within these external sources.
Capability inventory: The skill has the capability to read and write files to the local project directory and perform web searches, which could be leveraged if an injection is successful.
Sanitization: No sanitization, escaping, or validation of the content retrieved from local files or web research is performed before the agent processes it.
[EXTERNAL_DOWNLOADS]: The "Live Research Pass" feature utilizes built-in web search or external skills (like deep-research or watch-video) to retrieve current information and citations regarding marketing advisors, which involves downloading and interpreting external web content.
[NO_CODE]: This skill does not include any executable code, scripts, or binary assets. It consists entirely of Markdown instructions and reference data files.

marketing-council