marketing-ideas

Pass

Audited by Gen Agent Trust Hub on Sep 25, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTIONDATA_EXFILTRATION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill ingest surface includes scraped web pages, reviews, and customer emails (SKILL.md). It mitigates risks by explicitly instructing the agent to treat these sources as untrusted input and to ignore embedded instructions. The skill leverages project context and external connectors for data gathering (connectors.md), while sanitization is managed by requiring verification of time-sensitive claims and distinguishing evidence from assumptions.\n- [DATA_EXFILTRATION]: While the skill can interface with external marketing tools (e.g., Google Ads, Stripe), it contains strict prohibitions against hardcoding API keys and requires explicit user consent for any persistent writes, spending, or account changes (SKILL.md and references/connectors.md).
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 25, 2026, 09:29 PM