Seo Blog Writer

Fully automated SEO article writer. Give it a topic and domain — it handles everything: auto-discovers product context, researches keywords, analyzes competi...

This skill appears to do what it says (research, write, verify links) and has no install or required secrets — good. However, its runtime instructions ask the agent to spawn a subagent with 'full' security so it can run shell commands (curl/grep) and write files incrementally. Only grant that level of access if you trust the skill and the platform prompts you with a clear, limited scope (which directories, which commands). If you prefer less risk: run the skill in your main session (as the SKILL.md suggests) so you control any file/network operations, or deny subagent exec permissions and accept a reduced feature set (no curl-based link checks or automated disk writes). Also audit any product_context files you supply so they don't contain secrets, and test the skill on a throwaway output directory first.

Type: OpenClaw Skill Name: seo-blog-writer Version: 3.2.0 The skill requests 'full' security permissions for subagents in SKILL.md, justifying it for disk writes and curl-based link verification. While these capabilities are plausibly linked to the stated goal of automated SEO writing, the requirement for high-privilege execution and the implementation of a complex 'long-running task' model with incremental file I/O represent a significant attack surface. No explicit evidence of malicious intent (such as data exfiltration or backdoors) was found, but the broad permissions and potential for shell-based execution via curl warrant a suspicious classification.