generate-plot
Warn
Audited by Gen Agent Trust Hub on Mar 10, 2026
Risk Level: MEDIUMCOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The CLI fallback command
paperbanana plot --data <file> --intent "<intent>"executes in a shell using$ARGUMENTS[1]without sanitization. While$ARGUMENTS[0]is checked for shell metacharacters, the intent string is not, posing a command injection risk if the agent falls back to the CLI. - [PROMPT_INJECTION]: The skill processes untrusted external data, creating an indirect prompt injection surface where malicious data could influence agent behavior.
- Ingestion points: File content read via the
Readtool in the Procedure (Step 1). - Boundary markers: No delimiters or instructions are used to isolate the data content.
- Capability inventory: Includes
Bashcommand execution and MCP tool usage. - Sanitization: Sanitization is applied to the file path but not to the content of the data file.
Audit Metadata