Security audit
Security checks for vulnerabilities and agentic risk
This is a manually installed Claude Code status-line helper with some documentation gaps, but no hidden execution, exfiltration, or destructive behavior was found.
Before installing, confirm you are comfortable editing Claude Code settings and running a persistent local status-line shell script. Expect only branch, model, and changed-line information to work as written; the other displayed fields are placeholders unless you extend the script.
Skill reads from agent configuration directories (.claude/, .codex/, .gemini/). These directories may contain API keys, personal settings, and other credentials that the skill has no legitimate need to access.
在 ~/.claude/settings.json 中添加：
{
Skill reads from agent configuration directories (.claude/, .codex/, .gemini/). These directories may contain API keys, personal settings, and other credentials that the skill has no legitimate need to access.
在 ~/.claude/settings.json 中添加：
{
Skill reads from agent configuration directories (.claude/, .codex/, .gemini/). These directories may contain API keys, personal settings, and other credentials that the skill has no legitimate need to access.
}
get_model() {
model_name=$(grep -o '"model": *"[^"]*"' ~/.claude/settings.json 2>/dev/null | cut -d'"' -f4 | tr -d '\n')
[ -z "$model_name" ] && model_name="unknown"
}
The manifest description says the skill displays Git branch, model, context usage, output token count, cost, duration, and changed lines in real time. However, the actual script only derives Git branch, model, and changed lines, and hardcodes the other four fields as 'N/A', so the advertised behavior is not implemented.
The natural-language content of the skill, including its description and instructions, is presented only in Chinese. Under the policy, forcing a specific language without user opt-in or justification is a locale/language policy violation.
The prerequisites section explicitly says jq is needed for JSON parsing, creating the impression that the implementation depends on jq. In reality, the script parses ~/.claude/settings.json using grep and cut, so the documentation contradicts the implementation rather than merely omitting detail.
No suspicious patterns detected.