The Agent Skills Directory

[REMOTE_CODE_EXECUTION]: The skill instructions in SKILL.md suggest running npx -y @coldiq/mcp@latest to connect the vendor's MCP server. This command downloads and executes code from the npm registry at runtime. As this package is provided by the skill's author (Cold-IQ), it is documented as intended functionality.
[EXTERNAL_DOWNLOADS]: The skill references the Cold-IQ API endpoint at https://api.coldiq.com and the vendor's marketplace at https://coldiq.com/marketplace for data enrichment services.
[PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection (Category 8) because it facilitates the ingestion of untrusted data from external sources.
Ingestion points: Data enters the agent's context through tools like mcp__coldiq__extract_post_engagement (which fetches social media posts) and mcp__coldiq__fetch_page_content (which fetches company page announcements), as mentioned in SKILL.md.
Boundary markers: There are no explicit instructions or delimiters in the templates provided in references/first-touch.md to prevent the agent from executing instructions that may be maliciously embedded within the fetched external data.
Capability inventory: The skill utilizes various data enrichment and verification tools from the Cold-IQ MCP, but it does not include local scripts with dangerous capabilities like arbitrary shell execution or file system modification.
Sanitization: The skill does not describe any validation or sanitization of the external content before it is used to populate the email templates.

cold-email-templates-34