The Agent Skills Directory

[SAFE]: The skill implements a structured analytical protocol for marketing decision-making. It relies on internal reference documents for methodology and uses user-provided context to perform text synthesis without dangerous tool invocations.- [INDIRECT_PROMPT_INJECTION]: The skill exhibits an attack surface for indirect prompt injection as it processes external marketing data and briefs provided by the user.
Ingestion points: The skill ingests untrusted data through user-defined decision briefs, options, and the local .agents/marketing-context.md file.
Boundary markers: The protocol utilizes specific templates (e.g., Council review, Lens record) to provide structure, though it does not use explicit 'ignore instructions' delimiters.
Capability inventory: Capabilities are limited to text generation and synthesis; while the skill mentions researching facts, this is handled via standard platform tools if available.
Sanitization: There are no explicit sanitization or filtering mechanisms for the ingested data mentioned in the skill instructions.

marketing-council