ClawHub
Back to skill

Security audit

qianfan-deepresearch

Security checks across malware telemetry and agentic risk

Overview

This skill is a disclosed Baidu Qianfan DeepResearch wrapper that uses your API key to generate reports and return download links.

Install only if you are comfortable sending research topics and generated report content to Baidu Qianfan under your API key. Prefer QIANFAN_API_KEY or a secure secret mechanism instead of pasting keys into chat or command lines, review the query before running, and remember the workflow proceeds automatically without showing the provider's clarification or outline approval steps first.

SkillSpector

By NVIDIA
Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
Findings (6)

Missing User Warnings

Medium
Confidence
93% confidence
Finding
The documentation explicitly promotes automatically sending "跳过" for clarification and "确认" for outline approval, which removes meaningful user review at two decision points. In a research/report-generation workflow, this can cause the system to proceed with misunderstood scope, biased assumptions, or unsafe content generation without informed user consent.

Missing User Warnings

Medium
Confidence
93% confidence
Finding
The workflow explicitly documents automatically sending “跳过” for clarification and “确认” for outline approval, which removes user review from meaningful decision points. In a research/report-generation skill, this can cause the agent to proceed with incorrect scope, unintended topics, or sensitive requests without operator confirmation, increasing the chance of unauthorized or inaccurate external processing.

Missing User Warnings

Medium
Confidence
97% confidence
Finding
The example passes the API key directly on the command line, which commonly exposes secrets through shell history, process listings, audit logs, and agent execution traces. In an agent environment, command invocations are often logged centrally, so the risk of credential disclosure is materially higher than in a local ad hoc script.

Missing User Warnings

Low
Confidence
83% confidence
Finding
The documented stdout JSON includes direct report download URLs without warning that they may grant access to sensitive generated research content. In automated agent pipelines, stdout is frequently captured by logs, orchestration systems, or downstream tools, which can unintentionally leak those links beyond the intended recipient.

Unbounded Resource Access

Medium
Category
Excessive Agency
Content
headers=HEADERS,
        json=payload,
        stream=True,
        timeout=None,  # 不设整体超时，用空闲超时控制
    ) as resp:
        resp.raise_for_status()
        last_data_time = time.time()
Confidence
79% confidence
Finding
timeout=None

Unbounded Resource Access

Medium
Category
Excessive Agency
Content
headers=make_headers(api_key),
        json=payload,
        stream=True,
        timeout=None,  # 不设整体超时，通过 IDLE_TIMEOUT 控制
    ) as resp:
        resp.raise_for_status()
Confidence
90% confidence
Finding
timeout=None

VirusTotal

65/65 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

Detected: suspicious.exposed_secret_literal

File appears to expose a hardcoded API secret or token.

Critical
Code
suspicious.exposed_secret_literal
Location
scripts/deepresearch.py:31