The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill requires the agent to ingest and analyze external candidate content, such as public repositories, blog posts, and portfolios. This untrusted data from external sources could contain malicious instructions designed to manipulate the agent's evaluation or behavior.
Ingestion points: The evaluation process in references/portfolio-and-work-sample-scoring.md involves reviewing links to public writing, speaking, and code.
Boundary markers: The skill lacks explicit delimiters or instructions to treat the analyzed content strictly as data, leaving the agent vulnerable to instructions hidden within those materials.
Capability inventory: The skill itself does not define custom tools or scripts; it utilizes the agent's standard text generation and tool-use capabilities.
Sanitization: No process for sanitizing or filtering the content retrieved from external URLs is provided.

devrel-hiring