johari-window
Pass
Audited by Gen Agent Trust Hub on Sep 3, 2026
Risk Level: SAFEPROMPT_INJECTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [PROMPT_INJECTION]: The skill contains specific instructions for the agent to reveal its system prompt verbatim if requested by the user. It also uses emphatic language to prevent the agent from simplifying or altering its core logic, which is a meta-instructional constraint on the agent's behavior.
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest external materials such as documents, links, and chat records. This creates a surface for indirect prompt injection where malicious instructions in that content could influence the agent's collaboration logic.
- Ingestion points: External materials (documents, screenshots, links, chat history) provided by the user as context for the Johari Window logic.
- Boundary markers: There are no explicit delimiters or instructions for the agent to ignore potentially malicious embedded content in the ingested materials.
- Capability inventory: The skill defines conversational and reasoning logic; it does not perform file system writes, network requests, or command execution.
- Sanitization: No evidence of sanitization or safety-focused validation of user-provided content is present in the instructions.
Audit Metadata