pillar-pages
Pass
Audited by Gen Agent Trust Hub on Oct 10, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [SAFE]: The skill is designed to guide agents in creating SEO content strategies. It provides clear, legitimate methodologies based on established industry frameworks like HubSpot and Ahrefs.
- [COMMAND_EXECUTION]: The skill includes a local utility script,
scripts/check-output.py, which is used to validate that the agent's output meets length and keyword requirements. The script performs basic text processing on a local file and does not involve network access, sensitive file manipulation, or execution of untrusted code. - [INDIRECT_PROMPT_INJECTION]: The instructions suggest gathering information from external sources such as Reddit and Quora to identify common questions for topic clusters. While processing data from these platforms presents a theoretical surface for indirect prompt injection, the skill focus is on content planning, and it does not implement automated scraping or ingestion mechanisms that would escalate this risk.
Audit Metadata