The Agent Skills Directory

[Indirect Prompt Injection] (HIGH): This skill is vulnerable to indirect prompt injection because it processes untrusted external data (matrix files and residual inputs) while maintaining high-privilege capabilities.
Ingestion points: Processes .npy matrix files and user-provided numeric strings via CLI arguments in scripts/.
Boundary markers: There are no explicit delimiters or instructions to the agent to ignore potentially malicious content within data files.
Capability inventory: The skill allows the use of the Bash tool to execute Python scripts and the Write tool to modify the filesystem.
Sanitization: The provided documentation does not specify any sanitization or validation logic for the ingested data.
[Command Execution] (LOW): The skill naturally uses the Bash tool to run its internal numerical scripts. While this is the intended functionality, it provides a vector for escalation if the agent is manipulated via the data ingestion points mentioned above.

linear-solvers