blog-topic-research
PassAudited by ClawScan on May 13, 2026.
Overview
The skill appears to be a coherent blog-topic research workflow, but users should review any credential prompts or optional local script execution before allowing them.
This looks safe for researching public blog-topic demand. Before using it, make sure any file append targets only your intended backlog file, decline unexplained credential requests, and review any Python script before allowing the optional similarity-check step.
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.
If approved, the agent can change the specified backlog or CSV file.
The skill can modify a local file, but the instruction requires presenting results and asking the user first, making it disclosed and user-directed.
`--append-to <path>` - after presenting results, ask the user before appending accepted topics as JSON to the given path
Only provide a path you intend to update, and review the proposed topics before approving the append.
Running a local script could read local files or perform other actions depending on what script is used.
The optional shell-out is purpose-aligned for similarity checking, but the referenced script is not included in the provided artifact set, so its behavior cannot be reviewed here.
The skill is content-only: it does no scraping of its own. It drives the agent's `WebFetch` and `WebSearch` tools to fetch sources, and (optionally) shells out to a Python similarity script for the cannibalization step.
Before allowing the cannibalization step, confirm which Python script will run and review or provide it yourself.