Security audit
Security checks for vulnerabilities and agentic risk
The skill’s research purpose is coherent, but it asks the agent to install/update and run mutable external code from a full Git repository in the user’s home directory.
Install only if you trust Cue and the sensedeal/cue-skills repository as live code providers. Before running, consider pinning the repository to a known commit, reviewing the runner script, and confirming that reading ~/.cue/config.json and spending Cue credits are acceptable.
The file’s user-facing instructions and description are consistently presented only in Chinese, and there is no opt-in, alternative language option, or justification that the skill is intentionally region-specific. Under the stated policy, forcing a specific language without user choice is a natural-language policy violation.
The skill instructs the agent to clone or update an entire external Git repository into the local home directory before use. That materially expands trust and execution scope beyond the declared research purpose: a compromised upstream repo, malicious mirror, or unexpected post-clone contents could introduce arbitrary code or unsafe tooling into the agent environment.
The skill directs the agent to execute a local Python script obtained from an external repository, creating a direct path from untrusted remote content to local code execution. In an agent setting, this is especially dangerous because the script can access local files, credentials, network resources, and persistence locations while appearing to perform a benign research task.
