The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes potentially untrusted external data including keyword exports (Ahrefs, SEMrush), sales call transcripts, survey responses, and forum posts from Reddit and Quora to generate marketing strategies. This creates a surface where instructions embedded in these inputs could influence the agent's output.
Ingestion points: External data sources identified include keyword data exports, customer call transcripts, survey responses, and web search results from community forums (SKILL.md).
Boundary markers: The instructions do not define specific delimiters or instructions to ignore embedded commands within the ingested text.
Capability inventory: The agent's role is restricted to strategic analysis and planning (e.g., creating pillars and clusters). It does not have access to file-write operations, shell execution, or arbitrary network calls within this skill.
Sanitization: The skill lacks explicit instructions for sanitizing or filtering external content before analysis.
[SAFE]: The provided documentation for headless CMS selection (references/headless-cms.md) is purely informational, referencing well-known services and industry best practices for content modeling without security risks.

content-strategy