The Agent Skills Directory

[PROMPT_INJECTION]: The skill exhibits a surface for indirect prompt injection by ingesting untrusted data from LinkedIn profiles and posts (via mcp__crustdata__enrich_person_by_linkedin and mcp__crustdata__get_person_linkedin_posts in SKILL.md).
Ingestion points: External data enters the context through LinkedIn enrichment tools and web searches.
Boundary markers: Absent; external data is interpolated into email hooks without delimiters.
Capability inventory: The skill is limited to text generation and lacks capabilities for shell execution, file writing, or unauthorized network operations.
Sanitization: No explicit sanitization or filtering of the external data is performed before interpolation.
[EXTERNAL_DOWNLOADS]: The skill references blog posts and case studies from the vendor's official domain (inkeep.com). These are used as resources for generating follow-up email content and are documented as legitimate vendor resources.
[DATA_EXFILTRATION]: The skill accesses prospect-related data such as names, titles, and career history to facilitate deep personalization of outreach emails. This behavior is consistent with the skill's primary function and stated purpose.

cold-email