ClawHub
Back to skill

Security audit

Serbia Job Search

Security checks for vulnerabilities and agentic risk

Overview

The skill mostly does the disclosed job-search workflow, but its optional email feature stores a Gmail App Password in plaintext and renders external job data into email without escaping.

Review this before installing. The core job-search behavior is coherent and there is no evidence of deception or auto-applying, but do not enable email unless you are comfortable with a Gmail App Password being saved in plaintext in the skill directory. Prefer a dedicated sending account or a secret store, restrict file access, and revoke the app password if the config is shared. Enable cron only with an explicit time and confirm how to delete it later.

A.I.G

Vulnerability Patterns
  • Insecure Skill Coding PracticesFinds exploitable flaws such as hardcoded secrets or command injection
  • 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
Findings (2)

T09 · Insecure Skill Coding Practices

Error
Location
SKILL.md:143
Finding

Gmail App Password Stored in Plaintext Configuration

Content
View full analysis
Remediation
View remediation

T09 · Insecure Skill Coding Practices

Warning
Location
scripts/send_email.py:51
Finding

Unescaped External Job Data Injected into HTML Email Digests

Content
View full analysis
" + "".join( f"{t}" for t in tags[:8] ) + "" return f"""
{job.get('title','')}
{job.get('company','')}  ·  {job.get('location','')}
{tag_html}
{score}
{label}
Why: {jo ...[truncated 3495 chars]
Remediation
View remediation

SkillSpector

Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Rogue AgentSelf-Modification, Session Persistence
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • Behavioral ASTexec() Call, eval() Call, Dynamic Import
  • MCP Least PrivilegeUnderdeclared Capability, Wildcard Permission, Missing Permission Declaration
Findings (8)

Missing User Warnings

High
Category
Not specified by scanner
Confidence
97% confidence
Finding

The skill asks the user for a Gmail App Password and instructs storing SMTP credentials in workspace/config.json without a prominent warning about secret handling, storage risk, or safer alternatives. If that file is exposed, logged, synced, or later read by other tools, the mailbox can be abused to send mail or access connected services, turning a convenience feature into a credential-compromise risk.

Content

No source excerpt is available for this finding.

Session Persistence

Medium
Category
Rogue Agent
Confidence
60% confidence
Finding

Skill establishes unauthorized persistence across sessions via cron jobs, startup scripts, or state files. Session persistence allows an attacker to maintain access beyond the current interaction.

Content

Scanner excerpt · README.md (reported line 16)May include surrounding context.

1. Install

bash
mkdir -p ~/.openclaw/skills
cp -r job-search-belgrade ~/.openclaw/skills/

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
96% confidence
Finding

The README explicitly instructs users to place a Gmail App Password in workspace/config.json, which encourages storing a reusable credential in plaintext inside the skill directory. If that directory is exposed through backups, logs, version control, local compromise, or other skill/runtime access, the credential could be stolen and used to access email-sending capabilities.

Content

No source excerpt is available for this finding.

Undeclared Tool Scope

Medium
Category
MCP Least Privilege
Confidence
93% confidence
Finding

The skill invokes shell commands, writes files, and performs networked actions, but it does not declare any explicit tool scope or permission boundaries. That increases the chance the agent can execute higher-risk actions than a user expects, including modifying config, scheduling jobs, sending email, or scraping external sites without clear least-privilege constraints.

Content

No source excerpt is available for this finding.

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
95% confidence
Finding

The trigger phrases are broad enough to match ordinary job-search requests, which can cause the skill to activate unexpectedly in contexts where the user did not intend scraping, resume processing, file access, or email-related workflows. Over-broad invocation expands the attack surface by increasing accidental execution of privileged behaviors.

Content

No source excerpt is available for this finding.

subprocess module call

Medium
Category
Dangerous Code Execution
Confidence
70% confidence
Finding

subprocess module calls execute external commands. Without careful input validation, this enables command injection.

Content

Scanner excerpt · scripts/extract_resume.py (reported line 58)May include surrounding context.

python
# Attempt 3: pdftotext CLI
    try:
        import subprocess
        result = subprocess.run(
            ["pdftotext", "-layout", str(pdf_path), "-"],
            capture_output=True, text=True, timeout=15
        )

Vague Triggers

Low
Category
Not specified by scanner
Confidence
89% confidence
Finding

The phrase "schedule this" is underspecified and can be interpreted in many contexts, making it easier for the skill to initiate persistence-oriented behavior such as cron registration without sufficiently clear user intent. In a skill that can create recurring tasks, ambiguity around scheduling is a meaningful safety issue even if the action is not inherently malicious.

Content

No source excerpt is available for this finding.

Description-Behavior Mismatch

Low
Category
Not specified by scanner
Confidence
78% confidence
Finding

The script's documented and implemented behavior is to mark all discovered jobs as 'seen' in a local database so they are excluded from future runs. While deduplication is related to job search operation, this stateful persistence is not mentioned in the manifest description, creating a mild description-to-behavior gap.

Content

No source excerpt is available for this finding.

Static analysis

Detected: suspicious.exposed_secret_literal

File appears to expose a hardcoded API secret or token.

Critical
Code
suspicious.exposed_secret_literal
Location
README.md:68