skills/plurigrid/asi/modern-python/Gen Agent Trust Hub

modern-python

Warn

Audited by Gen Agent Trust Hub on Jun 22, 2026

Risk Level: MEDIUMEXTERNAL_DOWNLOADSREMOTE_CODE_EXECUTIONCOMMAND_EXECUTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill provides instructions to download installation scripts from external sources, specifically https://astral.sh/uv/install.sh and https://github.com/j178/prek/releases/latest/download/prek-installer.sh. These are used to set up project management and pre-commit tools.
  • [REMOTE_CODE_EXECUTION]: Recommends the use of piped shell execution patterns (curl | sh) for installing the uv and prek tools. This method executes remote scripts directly in the user's environment with high impact if the source or connection is compromised.
  • [COMMAND_EXECUTION]: The migration guide in SKILL.md provides a shell script snippet that iterates through a requirements.txt file and executes uv add for each entry. This represents an automated command execution surface that operates on external, untrusted data.
  • Ingestion points: requirements.txt file processed in the migration guide (SKILL.md).
  • Boundary markers: None; the script assumes valid package names but does not verify them against a whitelist or escape them.
  • Capability inventory: uv add (package installation and script execution via hooks).
  • Sanitization: None; simple grep filters are used to strip comments and empty lines but do not sanitize package identifiers.
  • [DATA_EXFILTRATION]: Includes example scripts in SKILL.md and references/pep723-scripts.md that perform network requests to httpbin.org and api.example.com. While intended for demonstration, these demonstrate the capability to transmit data to external domains.
Audit Metadata
Risk Level
MEDIUM
Analyzed
Jun 22, 2026, 06:42 AM
Security Audit — agent-trust-hub — modern-python