Back to skill

Security audit

storm-ready

Security checks for vulnerabilities and agentic risk

Overview

This skill is a local emergency-preparedness planner with disclosed local profile storage and no evidence of hidden network access, deception, or unsafe persistence.

Install only if you are comfortable storing household preparedness details in a local JSON profile. Use it as a planning aid, not for real-time emergency decisions; follow local emergency management and weather-service instructions during an active event.

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
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Behavioral ASTexec() Call, eval() Call, Dynamic Import
  • MCP Least PrivilegeUnderdeclared Capability, Wildcard Permission, Missing Permission Declaration
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
Findings (9)

Harmful Content Injection

Critical
Category
Prompt Injection
Confidence
95% confidence
Finding

This content may contain harmful instructions that could cause physical harm if followed. CRITICAL: Review carefully before use.

Content

Scanner excerpt · SKILL.md (reported line 3)May include surrounding context.

md
---
name: storm-ready
description: "Use when a hurricane/typhoon, ice storm, tornado outbreak, atmospheric river, wildfire season, or blizzard is approaching your area, when building a household emergency kit, when deciding how much water and food to store, when sizing a generator or power station for outages, when an advisory or watch is issued and you don't know what to do in what order, or annually to audit family preparedness — profiles your region's real hazards (gulf-coast hurricanes demand 14-day supplies, tornado country demands shelter kits, wildfire WUI demands go-bags), sizes water/food/power for YOUR people and pets with real math, runs a gap audit against your inventory, and prints phase-by-phase countdown checklists from T-72h through the storm to recovery, with the life-safety rules (CO, flood water, downed lines) that kill people every year."
version: 1.0.0
author: Denis Voronin
license: MIT
tags: [emergency, preparedness, hurricane, storm, outage, generator, water-storage, family-safety]
---

Undeclared Tool Scope

Medium
Category
MCP Least Privilege
Confidence
90% confidence
Finding

The skill advertises executable commands that invoke Python scripts, read/write a profile file under the user's home directory, and potentially use shell execution, but the manifest does not declare any tool scope or allowed-tools restrictions. This creates unnecessary ambiguity about what the skill is permitted to do and weakens least-privilege controls, increasing the chance of unintended file modification or command execution in agent environments.

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/test_storm_ready.py (reported line 103)May include surrounding context.

python
print("cli round-trip:")
with tempfile.TemporaryDirectory() as td:
    pf = os.path.join(td, "profile.json")
    r = subprocess.run([sys.executable, os.path.join(HERE, "storm_ready.py"),
                        "profile", "--file", pf, "--region", "ice-storm",
                        "--people", "3", "--pets", "dog:45"],
                       capture_output=True, text=True)

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/test_storm_ready.py (reported line 112)May include surrounding context.

python
saved = json.load(f)
    check("pets parsed", saved["pets"][0]["kind"] == "dog"
          and saved["pets"][0]["weight_lb"] == 45)
    r = subprocess.run([sys.executable, os.path.join(HERE, "storm_ready.py"),
                        "profile", "--file", pf, "--mark-have", "radio",
                        "first-aid"], capture_output=True, text=True)
    check("mark-have works", r.returncode == 0 and "2 marked have" in r.stdout)

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/test_storm_ready.py (reported line 116)May include surrounding context.

python
"profile", "--file", pf, "--mark-have", "radio",
                        "first-aid"], capture_output=True, text=True)
    check("mark-have works", r.returncode == 0 and "2 marked have" in r.stdout)
    r = subprocess.run([sys.executable, os.path.join(HERE, "storm_ready.py"),
                        "audit", "--file", pf], capture_output=True, text=True)
    check("audit shows ✓ for marked items", "✓" in r.stdout)
    check("audit flags P0 gaps", "P0 prep gaps" in r.stdout)

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/test_storm_ready.py (reported line 120)May include surrounding context.

python
"audit", "--file", pf], capture_output=True, text=True)
    check("audit shows ✓ for marked items", "✓" in r.stdout)
    check("audit flags P0 gaps", "P0 prep gaps" in r.stdout)
    r = subprocess.run([sys.executable, os.path.join(HERE, "storm_ready.py"),
                        "timeline", "--region", "wildfire-wui"],
                       capture_output=True, text=True)
    check("timeline renders phases", "T-72 h" in r.stdout and "after the storm" in r.stdout)

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/test_storm_ready.py (reported line 124)May include surrounding context.

python
"timeline", "--region", "wildfire-wui"],
                       capture_output=True, text=True)
    check("timeline renders phases", "T-72 h" in r.stdout and "after the storm" in r.stdout)
    r = subprocess.run([sys.executable, os.path.join(HERE, "storm_ready.py"),
                        "power", "--list"], capture_output=True, text=True)
    check("power --list shows loads", "oxygen-conc" in r.stdout)
    r = subprocess.run([sys.executable, os.path.join(HERE, "storm_ready.py"),

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/test_storm_ready.py (reported line 127)May include surrounding context.

python
r = subprocess.run([sys.executable, os.path.join(HERE, "storm_ready.py"),
                        "power", "--list"], capture_output=True, text=True)
    check("power --list shows loads", "oxygen-conc" in r.stdout)
    r = subprocess.run([sys.executable, os.path.join(HERE, "storm_ready.py"),
                        "example"], capture_output=True, text=True)
    check("example runs end-to-end", r.returncode == 0
          and "Preparedness audit" in r.stdout)

Missing User Warnings

Low
Category
Not specified by scanner
Confidence
84% confidence
Finding

This code stores user household details such as people, pets, medical devices, and inventory state in ~/.storm-ready.json. While the file path is disclosed, there is no explicit warning that potentially sensitive preparedness and household information will be persisted locally, which is relevant user-facing disclosure for a code skill that writes personal data.

Content

No source excerpt is available for this finding.