modal-compute

Pass

Audited by Gen Agent Trust Hub on Sep 8, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONREMOTE_CODE_EXECUTIONDYNAMIC_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill fetches the modal Python package from the official registry. modal is a well-known service for serverless compute.
  • [COMMAND_EXECUTION]: The skill uses the modal CLI to perform setup, open interactive shells, and execute scripts (e.g., modal run, modal setup, modal shell).
  • [REMOTE_CODE_EXECUTION]: The primary purpose of the skill is to transfer and execute local Python code on Modal's remote serverless GPU infrastructure.
  • [DYNAMIC_EXECUTION]: The skill involves creating and executing Python scripts (script.py) as part of the compute workflow.
  • [INDIRECT_PROMPT_INJECTION]: The skill creates a surface for indirect prompt injection as it facilitates the execution of scripts that may incorporate untrusted data.
  • Ingestion points: The script.py file processed by the modal run command.
  • Boundary markers: None provided to distinguish between instructions and data within the script.
  • Capability inventory: Remote execution of code on high-performance infrastructure via modal run.
  • Sanitization: No sanitization or validation of script contents is described.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 8, 2026, 01:46 AM
Security Audit — agent-trust-hub — modal-compute