The Agent Skills Directory

[Data Exposure Through Portable Artifacts]: The skill aggregates information from sensitive internal sources, including CRM systems (win/loss data), internal chat logs (Slack), and customer call transcripts, into a self-contained HTML file. While this functionality is central to its purpose, the suggestion to host this file on external third-party services like Netlify or Vercel introduces a potential risk of exposing proprietary sales intelligence and internal communications if access controls are not strictly managed.
[Untrusted Data Ingestion]: During the research phase, the skill fetches data from various public websites and review platforms. This introduces an attack surface where malicious content on those sites could attempt to influence the agent's output or bias the final analysis.
Ingestion points: Automated web searches targeting competitor product pages, blogs, and customer reviews (SKILL.md, Phase 3).
Boundary markers: The skill does not specify delimiters or instructions for the agent to ignore instructions found within the researched external content.
Capability inventory: The agent utilizes web search tools and has the ability to generate and save local HTML files (SKILL.md, Phase 5).
Sanitization: No explicit sanitization or filtering of the gathered web content is described prior to its inclusion in the generated battlecard artifact.

competitive-intelligence