The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes data from external sources, creating a surface for indirect prompt injection where malicious instructions could be embedded in news feeds or articles.
Ingestion points: The skill fetches external content from Google News RSS, Hacker News API, and Reddit (in references/newsjacking.md), as well as general web content via the dev-browser tool (in references/podcast-guest-prep.md).
Boundary markers: There are no explicit instructions or delimiters defined to isolate external data from the agent's primary instructions.
Capability inventory: The agent has capabilities for shell execution via curl in references/newsjacking.md, web browsing via dev-browser in references/podcast-guest-prep.md, and local file modification by writing to .agents/media-list.md in references/journalist-pitching.md.
Sanitization: The instructions do not specify any sanitization or validation steps for the data retrieved from external APIs or websites.
[COMMAND_EXECUTION]: The skill provides specific shell commands for the agent to execute as part of its research workflow.
Evidence: references/newsjacking.md includes bash snippets using curl, jq, and xmllint to retrieve and parse data from public APIs and RSS feeds.
[EXTERNAL_DOWNLOADS]: The skill retrieves data from several well-known external services.
Evidence: references/newsjacking.md fetches information from Google News (news.google.com), Hacker News (hn.algolia.com), and Reddit (reddit.com) to monitor trending topics and industry news.

public-relations