Pure Search

The name/description say it uses DuckDuckGo + trafilatura and returns cleaned markdown — the included script implements exactly that by requesting html.duckduckgo.com and using trafilatura.extract. Minor mismatch: SKILL.md suggests installing the 'duckduckgo-search' package, but the script scrapes DuckDuckGo HTML directly rather than importing that library; this appears to be a benign inconsistency (documentation vs implementation), not a capability mismatch.

SKILL.md is explicit: install trafilatura and use the script. The runtime instructions and script perform network requests to DuckDuckGo and to the target webpages and return the extracted markdown. This is within the stated purpose, but note that the skill will fetch arbitrary third-party pages and return their content (expected for a search/ extraction tool) — users should be aware it may download content from sites they query (including potentially private or malicious pages if given such URLs).

No install spec is embedded in the registry (instruction-only), so nothing is written to disk by the platform. SKILL.md asks the user to pip install dependencies (standard practice). The pip recommendation includes 'duckduckgo-search' even though the script does not use that package — a minor documentation mismatch but not an install-time risk from the registry itself.

The skill requests no environment variables, no credentials, and no config paths. The operations (HTTP GET to DuckDuckGo and to result URLs) do not require secrets; the requested privileges are proportional to the described functionality.

The skill is not configured as always: true and does not request persistent or cross-skill configuration. It can be invoked by the agent normally. Nothing here indicates elevated platform privileges or unexpected persistence.