The Agent Skills Directory

[SAFE]: The skill provides guidelines for extracting insights from standard customer research sources such as transcripts, surveys, and support tickets. The logic is focused on synthesis and categorization without involving dangerous execution patterns.
[SAFE]: All external references target well-known and trusted services including Reddit, G2, LinkedIn, Hacker News, and SparkToro. These are standard platforms for market research and are accessed via standard search operators or manual browsing.
[SAFE]: The use of project-specific context files (e.g., .agents/product-marketing.md) follows established best practices for agent state management and does not involve unauthorized access to sensitive system files.
[INDIRECT_PROMPT_INJECTION]: The skill inherently processes untrusted data from external research sources (reviews, forum posts, transcripts).
Ingestion points: Raw research material, community posts, and review site data referenced in SKILL.md and references/source-guides.md.
Boundary markers: The instructions focus on extraction and synthesis but do not explicitly warn the agent to disregard instructions found within the research data.
Capability inventory: The skill scripts do not contain subprocess calls, network write operations, or dynamic execution patterns.
Sanitization: No specific content sanitization is mentioned, though the analysis framework emphasizes clustering and thematic extraction which naturally filters noise.

customer-research