product-ad-shot-design-by-1oyim0r
Pass
Audited by Gen Agent Trust Hub on Oct 9, 2026
Risk Level: SAFE
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill references several external URLs for 'GoodCase' evidence and finished media examples (e.g.,
goodcase.ai,media.goodcase.ai). These are used for visual reference and prompt grounding. As these are established platforms for this niche and the skill author (learnprompt) is using them for the skill's primary purpose of design synthesis, they are considered safe references. - [COMMAND_EXECUTION]: The skill uses instructional language for the agent to use 'available image, video, or browser tools' to inspect media. This is a standard request for agents with multi-modal capabilities and does not involve arbitrary shell commands or privilege escalation.
- [DATA_EXFILTRATION]: No patterns for exfiltrating sensitive local data were found. The skill explicitly instructs the agent not to 'upload private assets' or 'make external account changes' without approval.
- [PROMPT_INJECTION]: The instructions contain phrases like 'IMPORTANT' or 'Required evidence gate', but these are used to define the skill's logic (workflow consistency) rather than attempting to bypass safety filters or override system instructions.
- [INDIRECT_PROMPT_INJECTION]: The skill ingests data from a local file (
references/cases.md). While this is an ingestion point, the content is static and provided as part of the skill package, posing no risk of third-party injection during standard execution.
Audit Metadata