concept-exercise-generator
Pass
Audited by Gen Agent Trust Hub on Aug 3, 2026
Risk Level: SAFE
Full Analysis
- [PROMPT_INJECTION]: No attempts to override safety guidelines or agent behavior were detected. The instructions are focused on educational content generation.
- [DATA_EXFILTRATION]: No network operations or sensitive file path access (e.g., .ssh, .env) were found. The skill only generates local markdown files.
- [OBFUSCATION]: No obfuscated content, Base64 encoding, or hidden characters were found.
- [REMOTE_CODE_EXECUTION]: The skill does not download or execute any external scripts or packages.
- [PRIVILEGE_ESCALATION]: No commands for acquiring higher permissions (e.g., sudo, chmod) are present.
- [PERSISTENCE]: No mechanisms for maintaining access across sessions (e.g., cron jobs, shell profiles) were identified.
- [METADATA_POISONING]: The metadata in the YAML frontmatter accurately describes the skill's functionality.
- [INDIRECT_PROMPT_INJECTION]: The skill ingests user-provided source material to generate exercises.
- Ingestion points: Processes 'source material' (papers, lectures, text) to create exercises.
- Boundary markers: None explicitly defined for input material.
- Capability inventory: No dangerous tools; output is restricted to generating markdown text.
- Sanitization: Not present, but risks are mitigated by the lack of dangerous capabilities.
- [TIME_DELAYED_ATTACKS]: No logic based on date, time, or specific triggers was found.
- [DYNAMIC_EXECUTION]: No use of dynamic code execution (e.g., eval, exec) or runtime compilation.
Audit Metadata