pmem-style-generation

Pass

Audited by Gen Agent Trust Hub on Mar 10, 2026

Risk Level: SAFE
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill installs the powermem and python-dotenv Python packages during its environment setup phase. These are functional dependencies provided by the vendor to enable memory and configuration management.
  • [COMMAND_EXECUTION]: The skill orchestrates several local Python scripts to handle its internal logic, including check_config.py for API key validation, match_profile.py for stylistic intent matching, and save_generation.py for file management. These operations are confined to the skill's directory and temporary folders.
  • [SAFE]: The skill follows security best practices for handling untrusted external data. It implements a two-step process: first, a 'neutral information extraction' prompt (dehydration) strips away stylistic or instructional noise from user-provided content using clear boundary markers (---). This extracted, objective summary is then used as the basis for the final generation, effectively mitigating the risk of indirect prompt injection.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 10, 2026, 08:30 AM
Security Audit — agent-trust-hub — pmem-style-generation