deepthink
Pass
Audited by Gen Agent Trust Hub on Oct 8, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill invokes a local Python script to orchestrate the reasoning workflow. The command
python3 -m skills.deepthink.think --step 1is used to initiate the structured analysis process. - [INDIRECT_PROMPT_INJECTION]: The skill is designed to process user-supplied analytical questions. Because it uses these inputs to drive a multi-step reasoning chain involving sub-agents, there is a theoretical surface for indirect prompt injection if the user input contains instructions designed to subvert the workflow logic. However, the skill implements structural phases (abstraction, characterization, and factored verification) which naturally serve as mitigating factors against simple prompt-based overrides.
Audit Metadata