learn-tech
Pass
Audited by Gen Agent Trust Hub on Feb 21, 2026
Risk Level: SAFE
Full Analysis
- PROMPT_INJECTION (SAFE): The skill uses a 'patient teacher' persona to structure its responses. No instructions to bypass safety filters or override system prompts were found.
- DATA_EXPOSURE & EXFILTRATION (SAFE): The skill writes output to a specific local directory (
learn-tech_outputs/). No access to sensitive system files or credentials was detected. No external network exfiltration patterns are present. - REMOTE_CODE_EXECUTION (SAFE): No external scripts are downloaded or executed. The reference to
@mermaid-generatoris a standard inter-skill call for diagramming. - COMMAND_EXECUTION (SAFE): No arbitrary shell command execution or privilege escalation attempts were found.
- INDIRECT_PROMPT_INJECTION (LOW): The skill ingests a user-provided topic. While this is a common surface for prompt injection if the user provides malicious text as the topic, the skill's capabilities (writing markdown and generating diagrams) pose minimal risk. No sanitization logic is explicitly defined for the input topic.
Audit Metadata