image-to-code
Pass
Audited by Gen Agent Trust Hub on Sep 10, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONMETADATA_POISONING
Full Analysis
- [SAFE]: The skill consists entirely of natural language instructions and metadata. It does not include any executable scripts, binaries, or automated network operations. The instructions focus on directing the AI's creative process and implementation strategy for web development tasks.
- [INDIRECT_PROMPT_INJECTION]: The skill instructs the agent to ingest and analyze visual data and extract text from images to inform code implementation. This creates a surface for indirect prompt injection where instructions embedded in images could influence the agent's logic. 1. Ingestion points: Generated or provided images analyzed in SKILL.md (Sections 9, 21). 2. Boundary markers: None identified. 3. Capability inventory: The skill allows generating code artifacts but lacks system-level command execution or network access. 4. Sanitization: None described.
- [METADATA_POISONING]: The skill metadata and license refer to an author name and upstream repository ('Leonxlnx') that does not match the platform-provided author context. While this inconsistency is noted, it does not represent a functional security risk or a deceptive attempt to bypass safety controls.
Audit Metadata