paperbanana
Pass
Audited by Gen Agent Trust Hub on Nov 2, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill instructs the agent to invoke shell commands via the paperbanana.cli module to generate or evaluate images.
- Evidence: Commands like
python -m paperbanana.cli generateandpython -m paperbanana.cli plotare defined inSKILL.mdandreferences/commands.md. - Mitigation: The skill specifically directs the agent to quote arguments and avoid interpolating user prose directly into commands to prevent shell injection.
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted external data (PDFs, text files, CSV/JSON) which is passed to the generation or evaluation backend.
- Ingestion points: Data is ingested via file paths provided to commands like
generate --inputandplot --data. - Boundary markers: There are no instructions for the agent to use delimiters or specific ignore-rules for content within these files.
- Capability inventory: The agent can read local files and execute the CLI tool which communicates with external AI providers.
- Sanitization: No explicit content sanitization is described for the data within the input files.
Audit Metadata