The Agent Skills Directory

[SAFE]: The skill consists entirely of instructional Markdown content and templates for hiring processes. It does not contain executable code, script files, or network operations. All resources are static text files designed to guide the agent's behavior during a manual evaluation process.
[NO_CODE]: No scripts or binaries are included with the skill. The logic is defined through natural language instructions and structured templates.
[DATA_EXFILTRATION]: The skill does not access sensitive local files or perform network operations. It explicitly encourages privacy by instructing users to provide redacted excerpts for sensitive candidate information and avoid requesting secrets.
[PROMPT_INJECTION]: The skill is designed to process external candidate data, which presents a surface for indirect prompt injection; however, the risk is minimal as the skill lacks dangerous capabilities. \n
Ingestion points: Candidate resumes, portfolio materials, and interview notes (SKILL.md). \n
Boundary markers: None explicitly defined in the prompts. \n
Capability inventory: The skill uses no tools for network access, file-system modification, or command execution. \n
Sanitization: No specific sanitization or filtering is performed on input data.

evaluating-candidates