skill-system-behavior

Pass

Audited by Gen Agent Trust Hub on Mar 11, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: The skill's primary function is to manage and validate behavior specifications for other agent skills, following a BDD (Behavior-Driven Development) workflow.
  • [COMMAND_EXECUTION]: The skill uses local Python scripts for its operations (e.g., validate_spec.py, verify_structural.py). These scripts are bundled with the skill and perform expected file-processing tasks related to the skill's stated purpose.
  • [EXTERNAL_DOWNLOADS]: The skill identifies a dependency on the standard PyYAML Python package for processing YAML files. No untrusted or non-standard external dependencies or remote code execution patterns were found.
  • [DATA_EXFILTRATION]: No network activity or attempts to access sensitive system files (such as credentials or SSH keys) were detected in the scripts or metadata. Data processing is confined to the local filesystem.
  • [PROMPT_INJECTION]: The instructions focus on technical workflow management and do not include patterns designed to bypass safety filters, override agent behavior, or extract system prompts.
  • [INDIRECT_PROMPT_INJECTION]: The skill ingests local project files and specifications. While this represents a data ingestion surface, the processing is limited to structural validation and documentation generation, with no interpolation of untrusted data into instructions that would influence agent behavior.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 11, 2026, 11:36 PM