marketing-context

Pass

Audited by Gen Agent Trust Hub on Aug 28, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
  • [SAFE]: The skill instructions establish clear boundaries for marketing work, including a prohibition on scraping restricted sources or storing sensitive personal data without authorization. The use of local file writing to .agents/marketing-context.md is a standard persistence mechanism for maintaining project context and does not involve risky privilege escalation or command execution.
  • [INDIRECT_PROMPT_INJECTION]: The skill possesses a surface for indirect prompt injection as it is designed to ingest and process data from external primary sources and customer research artifacts.
  • Ingestion points: External websites, product reviews, and interview transcripts are processed to build market context (SKILL.md).
  • Boundary markers: The instructions explicitly require the agent to distinguish between 'inspected facts, user assertions, analysis, hypotheses, and unknowns', providing a logical boundary against instructions hidden in external data (SKILL.md).
  • Capability inventory: The skill involves network browsing of primary sources and file system writes to a designated project context file (SKILL.md).
  • Sanitization: The skill instructions do not specify automated sanitization, relying instead on the agent's directive to verify facts and preserve verbatim language with source context.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 28, 2026, 11:55 AM
Security Audit — agent-trust-hub — marketing-context