Security audit
Security checks for vulnerabilities and agentic risk
This skill is a small Hugging Face model-listing helper with disclosed network access and user-chosen JSON export, with minor caveats around optional scheduled use and imperfect “trending” wording.
Install only if you are comfortable with the skill making Hugging Face API requests, optionally through a proxy you provide, and writing exports to paths you specify. Do not add the cron example unless you intentionally want recurring daily network access and log output.
The skill documentation describes network access to Hugging Face and JSON file export, but the skill metadata does not declare any explicit tool scope such as permissions or allowed-tools. This weakens sandboxing and review because consumers cannot tell up front that the skill may perform outbound network requests and write files.
The sample output is presented as Hugging Face 热门模型, indicating the skill may emit Chinese-language text by default. The file does not document any user choice or opt-in for language/locale, which conflicts with the policy against forcing a specific language without user selection.
The skill documentation recommends creating a cron job for daily execution, which establishes persistence outside the immediate user session. In an agent skill context, instructions that encourage scheduled background execution can lead to unintended long-lived activity, repeated network access, and ongoing data writes without fresh user approval.
Check trending models daily for new releases:
# Create cron job for daily monitoring
0 9 * * * cd /home/ltx/.openclaw/workspace && \
/home/ltx/.openclaw/workspace/skills/huggingface-trends/scripts/hf_trends.py \
-n 20 -p http://172.28.96.1:10808 >> /tmp/hf-trends.log 2>&1
The docstring says this function fetches trending models from the Hugging Face API, while the implementation comments admit there is no public trending endpoint and the code fetches generic model listings before sorting by likes. This is an active contradiction between documented intent and actual behavior, not merely an omitted detail.
The manifest centers on trending-model monitoring and comparison, but the code acknowledges it cannot access a public trending endpoint and substitutes a popularity proxy. That means the implemented behavior does not fully match the claimed capability of fetching trending models.
The header string hard-codes Chinese text ("热门模型" and "个") in the default output. This imposes a specific language on all users without any opt-in, language selection, or documented justification for a locale-specific behavior.
No suspicious patterns detected.