mirofish-prediction
Pass
Audited by Gen Agent Trust Hub on Sep 21, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted external data (news, market signals, policy changes) to build simulation narratives used in the agent's decision-making process. This creates a vulnerability to indirect prompt injection where malicious instructions in the source data could bias the output.
- Ingestion points: Seed data from news, market signals, and policy changes.
- Boundary markers: None specified to prevent the agent from following embedded instructions in external data.
- Capability inventory: Execution of local Python and Node.js applications, and LLM API calls via OpenRouter or Claude.
- Sanitization: No sanitization or filtering of external content is described.
- [COMMAND_EXECUTION]: The skill contains shell instructions to activate environments and run local Python and Node.js servers. While paths are static, executing local code requires users to ensure the source project is trustworthy.
Audit Metadata