The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process potentially untrusted data from multiple external and local sources, which could be exploited to influence the agent's behavior via embedded instructions.
Ingestion points: The skill reads project-specific context from .agents/product-marketing.md (or similar paths) and custom advisor dossiers from .agents/advisors/. It also incorporates data from a "Live Research Pass" using web search, deep-research, or watch-video skills.
Boundary markers: The skill lacks explicit instructions to use delimiters (e.g., XML tags or triple quotes) or "ignore embedded instructions" warnings when interpolating external findings or local file content into its prompts.
Capability inventory: The skill possesses the capability to read local project files and execute network-dependent tools (search and research agents).
Sanitization: There are no documented steps to sanitize, escape, or validate the content retrieved from external research or user-supplied dossier files before it is processed by the simulation logic.

marketing-council