The Agent Skills Directory

[COMMAND_EXECUTION]: The skill provides shell commands in references/newsjacking.md using curl, jq, and xmllint to fetch and parse data from public APIs. It also recommends installing jq using the Homebrew package manager.\n- [EXTERNAL_DOWNLOADS]: The skill fetches public RSS and JSON data from well-known services, including Google News, Reddit, and Algolia (Hacker News), to assist with media monitoring and research.\n- [PROMPT_INJECTION]: The skill processes untrusted data from external news and social platforms. While this represents a surface for indirect prompt injection, the risk is mitigated by the skill's specific focus on generating professional pitches and the lack of high-privilege capabilities.\n- [SAFE]: The skill follows standard agent practices for managing context by reading from and writing to files within the .agents/ directory. No unauthorized persistence, credential harvesting, or obfuscation was found.

public-relations