hai-idea
Pass
Audited by Gen Agent Trust Hub on Oct 5, 2026
Risk Level: SAFE
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to process user-provided ideas, creating a data ingestion surface. However, the risk is negligible as the skill only generates text analysis and does not possess capabilities for file writing, network access, or command execution.
- Ingestion points: User-supplied ideas and project descriptions provided via natural language prompts (SKILL.md).
- Boundary markers: The workflow instructions require the agent to restate the idea in one sentence and evaluate it against specific dimensions, which naturally separates the user input from the evaluation logic.
- Capability inventory: The skill only produces text output in Markdown format; no shell execution, file system modification, or network requests are present.
- Sanitization: The instructions specifically guide the agent to 'remove decoration and excitement' from the user's input, which serves as a basic form of input normalization.
Audit Metadata