buyer-intent-writer

Pass

Audited by Gen Agent Trust Hub on Aug 31, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSINDIRECT_PROMPT_INJECTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill documentation provides an installation command npx skills add nexscope-ai/ecommerce-seo-geo-skills. This command targets a package from the author's organization for execution.
  • [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process external content that is not controlled by the user, which could contain instructions aimed at manipulating the agent.
  • Ingestion points: The skill specifically requests customer reviews, support questions, and competitor copy as inputs for generating product descriptions (SKILL.md).
  • Boundary markers: The workflow defines a requirement to separate confirmed facts from hypotheses and unknown data, which helps mitigate accidental obedience to embedded instructions.
  • Capability inventory: The skill definitions are knowledge-based and do not include tools for file system modification, network exfiltration, or arbitrary code execution.
  • Sanitization: The workflow includes a mandatory verification step requiring the agent to check all performance and factual claims against the verified evidence before generating the final output.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 31, 2026, 05:39 AM
Security Audit — agent-trust-hub — buyer-intent-writer