weekly-agency-review
Pass
Audited by Gen Agent Trust Hub on Aug 11, 2026
Risk Level: SAFE
Full Analysis
- [PROMPT_INJECTION]: Analysis of the system prompt confirms it contains only legitimate educational instructions. There are no attempts to override agent behavior, extract system prompts, or bypass safety guidelines.- [DATA_EXFILTRATION]: The skill utilizes session data provided as input (e.g., retrieval quality, confidence levels) to provide feedback. There are no network calls or mechanisms to exfiltrate this information to external servers.- [COMMAND_EXECUTION]: The skill does not define or invoke any shell commands or external tools. It operates entirely as a text-based conversational interaction.- [EXTERNAL_DOWNLOADS]: No external URLs, scripts, or packages are requested or downloaded by the skill.- [OBFUSCATION]: The skill content is provided in plain text. No Base64, hex encoding, zero-width characters, or other obfuscation techniques were detected.
Audit Metadata