noxinfluencer
Pass
Audited by Gen Agent Trust Hub on Oct 29, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSINDIRECT_PROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill metadata specifies the installation of the
@noxinfluencer/cliNode.js package. This is a vendor-owned resource required for the skill to perform its intended creator intelligence and marketing operations via the command line. - [INDIRECT_PROMPT_INJECTION]: The skill has a surface for indirect prompt injection as it ingests untrusted data from external social media profiles (YouTube, TikTok, Instagram) through creator search and profile commands (e.g.,
creator search,creator profile). This data is then processed to assist in marketing workflows. The skill effectively mitigates this risk by explicitly forbidding the agent from drafting outreach copy and requiring user approval for all mutations, email tasks, and messaging operations. - Ingestion points: Influencer data fetched via
creatorandtiktokCLI command families (SKILL.md). - Boundary markers: Instructions define a "Verdict Framework" for evidence interpretation and require the agent to relate findings back to user-specified business objectives.
- Capability inventory: The skill can write files (
export download), send emails (email send), and send messages (message send) (references/marketing-ops.md). - Sanitization: The skill uses human-in-the-loop controls, requiring user approval for all material actions and content before execution.
Audit Metadata