gc-minimal-zine-poster-v0-3
Pass
Audited by Gen Agent Trust Hub on Aug 13, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill is composed entirely of instructional Markdown and metadata files, with no executable scripts, binaries, or external code dependencies.
- [INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted user content such as themes, articles, and photographs (ingestion points: SKILL.md). It leverages the platform's built-in image-generation capabilities (capability inventory). It implements strict boundary markers and sanitization protocols by instructing the agent to distinguish between stylistic elements and content, and explicitly forbids the reproduction of source text, brands, or watermarks (boundary markers and sanitization: SKILL.md, references/reference-analysis.md).
- [DATA_EXPOSURE_AND_EXFILTRATION]: There are no indicators of unauthorized data access or exfiltration; the skill only handles user-provided data for its primary creative purpose.
- [DYNAMIC_CONTEXT_INJECTION]: The skill does not utilize any shell command interpolation or dynamic context execution features.
Audit Metadata