Back to skill

Security audit

Jusou Search

Security checks for vulnerabilities and agentic risk

Overview

The scanners classify this skill as suspicious because its script may send an API key and search query to a destination set through an unrestricted URL, and the inspection was only partial.

Consider avoiding installation until the API destination and handling of your key are clear. Search queries may be sent to Jusou and upstream search providers, so avoid sending private information. The scan was partial: one of two components was scanned, so the evidence is incomplete.

Vulnerability Patterns
  • Skill Instruction HijackingAlters the agent's session goals or safety constraints when the skill loads
  • Agent Memory PoisoningWrites attacker-controlled rules into memory that affect later sessions
  • Remote Payload Retrieval and ExecutionFetches external code whose behavior can change after review
  • Embedded Malicious CodeShips malicious scripts inside the skill and executes them locally
  • Unauthorized Access and Privilege EscalationObtains permissions beyond the task's legitimate needs
Vulnerability Patterns
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
  • Output HandlingUnvalidated Output Injection, Cross-Context Output, Unbounded Output
  • Taint TrackingDirect Taint Flow, Variable-Mediated Taint Flow, Credential Exfiltration Chain
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
Findings (3)

Tainted flow: 'req' from os.environ.get (line 52, credential/environment) → urllib.request.urlopen (network output)

Critical
Category
Data Flow
Confidence
93% confidence
Finding

Sending the API key as a Bearer token to the intended HTTPS service is legitimate authentication, not inherently exfiltration. However, JUSOU_BASE_URL accepts an arbitrary destination, including plaintext HTTP, and the request attaches the key without origin validation; an attacker able to influence that setting could capture the credential and query content.

Content

Scanner excerpt · scripts/jusou_search.py (reported line 57)May include surrounding context.

python
data=json.dumps(body).encode(),
        headers={"Authorization": "Bearer " + key, "Content-Type": "application/json"})
    try:
        with urllib.request.urlopen(req, timeout=a.timeout) as r:
            out = json.loads(r.read())
    except urllib.error.HTTPError as e:
        try:

Autonomous Decision Making

Medium
Category
Excessive Agency
Confidence
80% confidence
Finding

Skill enables autonomous high-impact decisions without human-in-the-loop verification. Critical operations (destructive commands, financial transactions, data deletion) should require explicit user confirmation.

Content

Scanner excerpt · SKILL.md (reported line 22)May include surrounding context.

md
## Setup

Set the environment variable `JUSOU_API_KEY` to the user's key from the console at https://jusouai.cn (keys look like `jsk_...`).
**New users get 100 free points on sign-up (valid 7 days)** — enough to try every engine. If there is no key, ask the user to sign up or redeem a card at jusouai.cn and put the key in the environment variable. Never ask the user to paste the key into the chat.

## How to search

The search workflow lacks a privacy warning about sending query content to external services.

Medium
Category
Output Handling
Confidence
90% confidence
Finding

The skill instructs the agent to submit searches through Jusou's API and selected search engines, but does not warn that query content leaves the user's environment. Queries derived from user requests could contain private information, and neither language version instructs the agent to remove sensitive content before transmission.

Content

No source excerpt is available for this finding.