Seo Blog Writer
ReviewAudited by ClawScan on May 10, 2026.
Overview
The SEO writing workflow is mostly coherent, but it asks the agent to run with full execution privileges for a task that only appears to need web research and scoped file output.
Treat this as a Review item rather than clearly malicious. If you use it, avoid granting full security unless you truly need shell-based checks, run it in a sandbox or dedicated folder, set the output directory explicitly, and review the plan before the hands-off writing phase starts.
Findings (2)
Artifact-based informational review of SKILL.md, metadata, install specs, static scan signals, and capability signals. ClawScan does not execute the skill or run runtime probes.
Running this as written could give the skill broad local command and file access while it performs automated research and output generation.
The skill directs use of full subagent execution privileges for a writing and link-checking workflow. The artifacts do not clearly limit which shell commands may run, where files may be written, or which URLs may be fetched.
sessions_spawn(... runtime: "subagent", security: "full" ... needed for disk writes and curl-based link checks)
Run it only in a restricted workspace, set a dedicated output directory, and require review/approval before granting full exec or running shell-based link checks.
The skill may continue through multiple phases and create persistent local logs and drafts until the task is complete.
The skill discloses long-running, resumable execution with persistent task files. This is aligned with its multi-phase writing purpose, but users should be aware that it will leave local state behind.
Long-running task mode. This skill generates a multi-phase artifact. Before Phase 0: create a `task-logs/{slug}/` workspace, write each phase output to disk incrementallyUse a project-specific folder, monitor the run, and delete task logs or generated drafts when they are no longer needed.