modular-skills

Pass

Audited by Gen Agent Trust Hub on Mar 5, 2026

Risk Level: SAFEPROMPT_INJECTIONCOMMAND_EXECUTIONEXTERNAL_DOWNLOADS
Full Analysis
  • [PROMPT_INJECTION]: The skill provides templates in modules/enforcement-patterns.md using phrases such as 'YOU MUST' and 'NON-NEGOTIABLE' to enforce behavior. These serve as prompt injection markers that could inadvertently override agent safety or operational boundaries. Additionally, the skill possesses an indirect prompt injection surface as it is designed to analyze external skill files. Evidence: 1. Ingestion points: skill-analyzer and module_validator process file paths (README.md). 2. Boundary markers: Absent. 3. Capability inventory: Python scripts execute line counting and complexity analysis (guide.md). 4. Sanitization: Absent.
  • [COMMAND_EXECUTION]: The framework documentation directs the execution of local shell commands and Python scripts for validation and token usage estimation, which involves command line interaction.
  • [EXTERNAL_DOWNLOADS]: The documentation mentions using pip install to manage package dependencies from external registries like PyPI.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 5, 2026, 03:42 PM