65/100 (Medium)

OpenClaw： suspicious
VirusTotal： benign
StaticScan： clean

OpenClaw： suspicious

The package mostly matches an EBE guardrail and an evaluation tool, but there are notable inconsistencies (duplicate/overlapping skills, mismatched paths/metadata, and undeclared LLM key usage) and pr...

[内容已截断]

VirusTotal： benign VT 报告

静态扫描： clean

No suspicious patterns detected.

README 未提供

{
    "latestVersion": {
        "_creationTime": 1774127866196,
        "_id": "k978pvdkkcvnsh6wz97tz1973h83b9gv",
        "changelog": "**Major update: Initial release of the agent-sentinel skill (Evaluate-before-Execute guardrail for OpenClaw agents).**\n\n- Introduces a mandatory evaluation layer (\"agent-sentinel\") enforcing safety and compliance before any high-stakes tool call.\n- Implements strict GO \/ NO-GO decisions (ALLOW, BLOCK, ADVISE) for critical actions—booking, payment, web search, and shell commands—based on user and safety policy specified in SENTINEL_CONFIG.md.\n- Provides a clear command-line interface and response schema; emits structured JSON decisions with severity and suggested alternatives.\n- Enforces robust child-safety, budget, and travel preferences via a configurable policy file.\n- Includes explicit user override and advisory protocols for handling preference and policy violations.\n- Complete documentation of configuration, invocation, and decision handling in the new SKILL.md.",
        "changelogSource": "auto",
        "createdAt": 1774127866196,
        "parsed": {
            "clawdis": {
                "envVars": [
                    {
                        "name": "ANTHROPIC_API_KEY",
                        "required": true
                    },
                    {
                        "name": "OPENAI_API_KEY",
                        "required": true
                    }
                ],
                "requires": {
                    "bins": [
                        "python3"
                    ]
                }
            }
        },
        "version": "1.0.3"
    },
    "owner": {
        "_creationTime": 0,
        "_id": "publishers:missing",
        "displayName": "sanjeet-toosi",
        "handle": "sanjeet-toosi",
        "image": "https:\/\/avatars.githubusercontent.com\/u\/139010128?v=4",
        "kind": "user",
        "linkedUserId": "kn78sbsedk93325qe653mg048d83agw3"
    },
    "ownerHandle": "sanjeet-toosi",
    "skill": {
        "_creationTime": 1774111209711,
        "_id": "kd72q24sre019k35ca8tmv5yhd83b57f",
        "badges": [],
        "createdAt": 1774111209711,
        "displayName": "Vigilance",
        "latestVersionId": "k978pvdkkcvnsh6wz97tz1973h83b9gv",
        "ownerUserId": "kn78sbsedk93325qe653mg048d83agw3",
        "slug": "vigilance",
        "stats": {
            "comments": 0,
            "downloads": 43,
            "installsAllTime": 1,
            "installsCurrent": 1,
            "stars": 0,
            "versions": 4
        },
        "summary": "Evaluate-before-Execute (EBE) guardrail for OpenClaw agents. Issues a mandatory GO \/ NO-GO decision before any high-stakes tool call. Enforces child-safety p...",
        "tags": {
            "latest": "k978pvdkkcvnsh6wz97tz1973h83b9gv"
        },
        "updatedAt": 1774128706603
    }
}