The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill instructions require the agent to ingest and analyze external data (such as LinkedIn posts, company news, and website changes) to personalize outbound emails. This creates a potential attack surface where malicious instructions could be embedded in the processed content to influence the agent's behavior.
Ingestion points: External research signals from LinkedIn, company news, and public websites (referenced in SKILL.md and personalization.md).
Boundary markers: The skill does not define technical delimiters for untrusted data, although it provides a "Quality Check" framework and instructions to avoid AI-generated patterns.
Capability inventory: The agent is tasked with generating email text and is instructed on how to handle replies via webhook integrations (mentioned in the Tool Integrations section).
Sanitization: Human review is recommended for Tier 1 and Tier 2 outreach efforts to ensure quality and relevance before sending.
[EXTERNAL_DOWNLOADS]: The skill contains multiple links to an external GitHub repository (coreyhaines31/marketingskills) for tool integration guides and registry information. These references are for documentation purposes and point to markdown files rather than executable scripts or packages.

cold-email