The Agent Skills Directory

[Indirect Prompt Injection] (HIGH): The skill is vulnerable to indirect prompt injection due to its core functionality. Ingestion points: External data enters the agent context via <project-path>, <target-path>, and the current directory (.) passed to scanning scripts in SKILL.md. Boundary markers: None are specified to protect the agent from instructions embedded within the analyzed code or configuration files. Capability inventory: The skill executes local Python scripts, runs package managers (npm, pip), and manages infrastructure via docker and kubectl. Sanitization: There is no evidence of input validation or instruction filtering.
[Unverifiable Dependencies] (MEDIUM): The workflow requires running npm install and pip install -r requirements.txt without specifying package versions or verified sources, which facilitates supply chain attacks.
[Command Execution] (MEDIUM): The skill performs frequent command-line operations (Docker, Kubernetes, Python scripts). If the paths provided to these tools are not strictly sanitized, it could lead to arbitrary command execution on the host system.

senior-secops