prototype
Pass
Audited by Gen Agent Trust Hub on Oct 26, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest data from research artifacts (e.g., '/research' output) and user prompts to generate executable code such as tests, TUI apps, or UI components. This creates a surface where malicious instructions embedded in the input data could influence the behavior of the generated code.
- Ingestion points: The skill specifically references reading assumptions from
/researchartifacts and processing user requests like 'prototype this' or 'spike this' (SKILL.md). - Boundary markers: The skill mitigates risk by requiring the agent to state the question and expected verdicts before writing code, providing logical isolation (FEASIBILITY.md).
- Capability inventory: The skill utilizes the agent's ability to write files and execute code via project-standard task runners such as pnpm, python, and bun (SKILL.md).
- Sanitization: The instructions do not specify manual sanitization of the input data, relying on the agent's core safety protocols.
Audit Metadata