website-lead-magnet
Pass
Audited by Gen Agent Trust Hub on Oct 11, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill operates entirely as a set of instructional prompts for text generation. It does not contain any executable scripts, command-line operations, or network exfiltration logic.
- [SAFE]: Includes explicit guardrails that prohibit the generation of deceptive marketing content, such as fake scarcity, invented testimonials, or misleading conversion statistics. It instructs the agent to use placeholders like '[PROOF NEEDED: ...]' for any unverifiable claims.
- [SAFE]: References numerous official regulatory and government domains (e.g., .fgov.be, .europa.eu, .vlaanderen.be) to ensure compliance with GDPR and consumer protection laws. These external references are handled as informational context for the agent and do not involve remote code execution.
- [SAFE]: Uses a local file ingestion pattern (reading '.agents/business-context.md') to scope the agent's knowledge to the specific business project. This follows established best practices for agent context management and does not involve accessing sensitive system credentials or environment variables.
- [INDIRECT_PROMPT_INJECTION]: The skill reads external project data from 'business-context.md'. However, the risk is mitigated by the skill's limited capabilities (text generation only) and instructions to use placeholders and validation sweeps to verify all output, preventing the blind execution of embedded instructions from untrusted data.
Audit Metadata