creative-problem-solver
Pass
Audited by Gen Agent Trust Hub on Oct 9, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [SAFE]: No malicious code, obfuscation, or unauthorized data access patterns were identified. The skill's instructions are focused on cognitive processing and strategic output within the conversation context.
- [INDIRECT_PROMPT_INJECTION]: The skill defines a clear ingestion surface for user-defined objectives and constraints. However, as the skill lacks any operational tools (file, network, or command execution), there is no pathway for data-driven instructions to cause harm, and the surface is considered safe.
- [COMMAND_EXECUTION]: The skill contains explicit instructions to avoid repository mutation and implementation, delegating such tasks to specialized execution owners. This prevents the agent from being coerced into performing unauthorized system changes during the strategy generation phase.
Audit Metadata