The Agent Skills Directory

[SAFE]: The skill is entirely composed of Markdown instructions and reference data. It does not include any scripts, executables, or external dependencies.
[DATA_EXPOSURE]: The skill instructs the agent to check for specific local context files (e.g., .agents/product-marketing.md or .claude/product-marketing.md) to provide better context for writing. This is a documented and standard practice for context-aware agents and does not target sensitive system files or credentials.
[PROMPT_INJECTION]: The instructions focus on improving the quality and human-like tone of generated emails. There are no patterns suggesting attempts to bypass safety filters, extract system prompts, or override agent constraints.
[REMOTE_CODE_EXECUTION]: No remote code execution patterns, network downloads, or package installations were found in any of the analyzed files.
[INDIRECT_PROMPT_INJECTION]: The skill ingests user-provided data (prospect roles, company details, value propositions) to generate email drafts. While this represents a standard data ingestion surface, the lack of dangerous capabilities (like shell execution or network writes) makes the risk negligible.

cold-email