The Agent Skills Directory

[SAFE]: The skill consists entirely of text-based guidelines for public relations and press release creation. It contains no scripts, binary files, or automated commands.
[DATA_EXPOSURE_AND_EXFILTRATION]: The skill suggests reading project-specific context files (such as .claude/project-context.md) to align PR content with brand positioning. This is a standard context-gathering behavior for project-based AI agents and does not involve accessing sensitive system credentials (like SSH keys or environment secrets) or exfiltrating data to external servers.
[INDIRECT_PROMPT_INJECTION]: The skill has a surface for processing project documentation. However, it lacks any high-privilege capabilities (no shell command execution, no network transmission, no file modification), making the impact of any potential indirect injection through project files negligible.

public-relations