ClawHub
Back to skill

Security audit

Academic Paper Finder

Security checks for vulnerabilities and agentic risk

Overview

This skill does what it advertises: searches PubMed, enriches paper records, writes RIS files, and can add user-selected papers to Zotero using the user's Zotero credentials.

Install only if you are comfortable sending search terms, DOIs, PMIDs, and paper metadata to PubMed, OpenAlex, and Zotero. Use a Zotero API key with the minimum needed permissions, avoid sharing logs that may mention item IDs or errors, and review batch imports before running them because they can add multiple items to your Zotero library.

A.I.G

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

SkillSpector

Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Taint TrackingDirect Taint Flow, Variable-Mediated Taint Flow, Credential Exfiltration Chain
  • MCP Least PrivilegeUnderdeclared Capability, Wildcard Permission, Missing Permission Declaration
  • MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
Findings (16)

Tainted flow: 'url' from os.environ.get (line 83, credential/environment) → urllib.request.urlopen (network output)

Critical
Category
Data Flow
Confidence
90% confidence
Finding

Credentials or environment variables flow to a network sink. This is a high-confidence indicator of credential exfiltration.

Content

Scanner excerpt · scripts/batch_import.py (reported line 23)May include surrounding context.

python
"""Fetch paper metadata from PubMed"""
    url = f"https://eutils.ncbi.nlm.nih.gov/entrez/eutils/efetch.fcgi?db=pubmed&id={pmid}&retmode=xml"
    
    with urllib.request.urlopen(url, timeout=30) as response:
        content = response.read().decode('utf-8')
    
    title_match = re.search(r'<ArticleTitle>([^<]+)</ArticleTitle>', content)

Tainted flow: 'req' from os.environ.get (line 85, credential/environment) → urllib.request.urlopen (network output)

Critical
Category
Data Flow
Confidence
90% confidence
Finding

Credentials or environment variables flow to a network sink. This is a high-confidence indicator of credential exfiltration.

Content

Scanner excerpt · scripts/batch_import.py (reported line 91)May include surrounding context.

python
req.add_header('Content-Type', 'application/json')
    
    try:
        with urllib.request.urlopen(req) as response:
            data = json.load(response)
            return data.get('successful', {}) != {}
    except:

Tainted flow: 'url' from os.environ.get (line 114, credential/environment) → urllib.request.urlopen (network output)

Critical
Category
Data Flow
Confidence
90% confidence
Finding

Credentials or environment variables flow to a network sink. This is a high-confidence indicator of credential exfiltration.

Content

Scanner excerpt · scripts/zotero_add.py (reported line 22)May include surrounding context.

python
"""Fetch paper metadata from PubMed"""
    url = f"https://eutils.ncbi.nlm.nih.gov/entrez/eutils/efetch.fcgi?db=pubmed&id={pmid}&retmode=xml"
    
    with urllib.request.urlopen(url, timeout=30) as response:
        content = response.read().decode('utf-8')
    
    # Parse XML

Tainted flow: 'url' from os.environ.get (line 114, credential/environment) → urllib.request.urlopen (network output)

Critical
Category
Data Flow
Confidence
90% confidence
Finding

Credentials or environment variables flow to a network sink. This is a high-confidence indicator of credential exfiltration.

Content

Scanner excerpt · scripts/zotero_add.py (reported line 65)May include surrounding context.

python
def search_doi_to_pmid(doi):
    """Convert DOI to PMID"""
    url = f"https://eutils.ncbi.nlm.nih.gov/entrez/eutils/esearch.fcgi?db=pubmed&term={urllib.parse.quote(doi)}&retmode=json"
    with urllib.request.urlopen(url) as response:
        data = json.load(response)
        results = data.get('esearchresult', {}).get('idlist', [])
        return results[0] if results else None

Tainted flow: 'req' from os.environ.get (line 116, credential/environment) → urllib.request.urlopen (network output)

Critical
Category
Data Flow
Confidence
90% confidence
Finding

Credentials or environment variables flow to a network sink. This is a high-confidence indicator of credential exfiltration.

Content

Scanner excerpt · scripts/zotero_add.py (reported line 84)May include surrounding context.

python
req.add_header('Zotero-API-Version', '3')
    
    try:
        with urllib.request.urlopen(req) as response:
            data = json.load(response)
            return len(data) > 0
    except:

Tainted flow: 'req' from os.environ.get (line 116, credential/environment) → urllib.request.urlopen (network output)

Critical
Category
Data Flow
Confidence
90% confidence
Finding

Credentials or environment variables flow to a network sink. This is a high-confidence indicator of credential exfiltration.

Content

Scanner excerpt · scripts/zotero_add.py (reported line 122)May include surrounding context.

python
req.add_header('Zotero-API-Version', '3')
    
    try:
        with urllib.request.urlopen(req) as response:
            data = json.load(response)
            return len(data) > 0
    except:

Tp4

High
Category
MCP Tool Poisoning
Confidence
95% confidence
Finding

If the skill claims broad research and export functionality but does not implement it, users may rely on it for data handling and publication workflows under false assumptions. In a skill that touches external APIs, citations, and file generation, this kind of misrepresentation increases operational and privacy risk because behavior cannot be accurately predicted from the manifest.

Content

No source excerpt is available for this finding.

Tp4

High
Category
MCP Tool Poisoning
Confidence
95% confidence
Finding

If the skill claims broad research and export functionality but does not implement it, users may rely on it for data handling and publication workflows under false assumptions. In a skill that touches external APIs, citations, and file generation, this kind of misrepresentation increases operational and privacy risk because behavior cannot be accurately predicted from the manifest.

Content

No source excerpt is available for this finding.

Tp4

High
Category
MCP Tool Poisoning
Confidence
98% confidence
Finding

If the skill claims broad research and export functionality but does not implement it, users may rely on it for data handling and publication workflows under false assumptions. In a skill that touches external APIs, citations, and file generation, this kind of misrepresentation increases operational and privacy risk because behavior cannot be accurately predicted from the manifest.

Content

No source excerpt is available for this finding.

Undeclared Tool Scope

Medium
Category
MCP Least Privilege
Confidence
80% confidence
Finding

The skill advertises commands that use environment variables, network access, and file output, but it does not declare any explicit tool scope or permissions boundaries. This is dangerous because reviewers and execution frameworks cannot easily enforce least privilege, increasing the chance of unintended credential exposure, uncontrolled network calls, or file writes when the skill is invoked.

Content

No source excerpt is available for this finding.

External Transmission

Medium
Category
Data Exfiltration
Confidence
81% confidence
Finding

This is a real external transmission to a third-party service not strictly required for the primary PubMed search function. In the skill context, the behavior is related to a legitimate feature, but it still increases privacy and supply-chain exposure because user-derived identifiers are shared with OpenAlex without explicit consent or service-boundary transparency.

Content

Scanner excerpt · scripts/pubmed_search.py (reported line 82)May include surrounding context.

python
# OpenAlex API - free, no key needed
    for pmid in pmids:
        try:
            url = f"https://api.openalex.org/works?filter=pmid:{pmid}"
            req = urllib.request.Request(url)
            req.add_header('User-Agent', 'OpenClaw/1.0')

Missing User Warnings

Low
Category
Not specified by scanner
Confidence
84% confidence
Finding

The setup instructions require a Zotero API key but provide no warning about secure storage, least-privilege key creation, or avoiding disclosure in logs and shared environments. While not an exploit by itself, poor credential-handling guidance increases the chance that users expose sensitive API tokens during setup or troubleshooting.

Content

No source excerpt is available for this finding.

Missing User Warnings

Low
Category
Not specified by scanner
Confidence
90% confidence
Finding

This code sends the user-provided DOI to the NCBI PubMed API, which is a network operation involving user data. While the function docstring describes the search purpose, there is no runtime disclosure or warning that input will be transmitted to an external service.

Content

No source excerpt is available for this finding.

Missing User Warnings

Low
Category
Not specified by scanner
Confidence
91% confidence
Finding

The script transmits title or keyword queries to PubMed via HTTP requests. Although this is functionally necessary for the skill, the code does not provide a clear user-facing warning or disclosure that entered search text leaves the local system.

Content

No source excerpt is available for this finding.

Missing User Warnings

Low
Category
Not specified by scanner
Confidence
91% confidence
Finding

This function sends author names and optional year filters to PubMed over the network. The operation lacks a visible notice to users that personally entered search terms are transmitted to a third-party API.

Content

No source excerpt is available for this finding.

Missing User Warnings

Low
Category
Not specified by scanner
Confidence
93% confidence
Finding

The script sends retrieved PubMed IDs to a second external service, OpenAlex, to obtain citation counts, but this secondary disclosure is not made explicit to the user. Even though PMIDs are public identifiers, forwarding user-derived search results to an additional third party expands data sharing and metadata exposure beyond the primary PubMed lookup.

Content

No source excerpt is available for this finding.

Static analysis

No suspicious patterns detected.