The Agent Skills Directory

[Indirect Prompt Injection] (HIGH): The skill possesses a significant attack surface by processing untrusted external data (user resumes and job descriptions) alongside high-privilege capabilities. An attacker could embed malicious instructions within a job description that the agent is 'analyzing' or 'using for tailoring,' potentially leading to unauthorized file access or malicious command execution.\n
Ingestion points: Processes external text via Review the provided resume content and Parse job description in SKILL.md.\n
Boundary markers: No delimiters (e.g., XML tags or triple quotes) or 'ignore embedded instructions' warnings are specified in the workflow.\n
Capability inventory: Executes shell commands via python3 scripts/generate_resume_pdf.py and performs file-write operations to /mnt/user-data/outputs/.\n
Sanitization: No evidence of input validation or sanitization before passing data to scripts or using it in decision-making.\n- [Command Execution] (MEDIUM): The skill explicitly instructs the agent to execute a Python script (scripts/generate_resume_pdf.py) using a shell command. If the arguments (e.g., the JSON input) are constructed using unsanitized user content, it could lead to command injection or argument manipulation.\n- [Dynamic Execution] (HIGH): The 'Code Style' section provides instructions for 'generating Python scripts for PDF creation.' Generating executable code at runtime based on untrusted user data (like resume content) is a critical security risk that can lead to arbitrary code execution if the agent interpolates user input directly into the generated script's logic.

resume-optimizer