Back to skill

Security audit

DHMZ Weather

Security checks for vulnerabilities and agentic risk

Overview

This skill is a straightforward Croatian weather lookup helper with minor privacy and control caveats around inferred locations.

Install only if you are comfortable with the agent using a mentioned or inferred city for convenience when you invoke the skill. For privacy, provide the city explicitly, such as /dhmz-weather Zagreb, so the agent does not rely on conversation context.

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
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
  • Rogue AgentSelf-Modification, Session Persistence
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
Findings (6)

Anti-Refusal Statement

High
Category
Anti-Refusal
Confidence
80% confidence
Finding

Skill instructs the agent to omit warnings, disclaimers, or ethical commentary. Stripping safety caveats hides risk from the user and is a common jailbreak preamble.

Content

Scanner excerpt · SKILL.md (reported line 62)May include surrounding context.

md
### Alerts
| Level | Emoji |
|-------|-------|
| Green (no warning) | 🟢 |
| Yellow | 🟡 |
| Orange | 🟠 |
| Red | 🔴 |

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 22)May include surrounding context.

Copy the SKILL.md file to your Claude Code skills directory:

bash
# Create skills directory if it doesn't exist
mkdir -p ~/.claude/skills/dhmz-weather

# Copy the skill file

Skill Enumeration

Medium
Category
Agent Snooping
Confidence
85% confidence
Finding

Skill enumerates or reads other installed skills. Access to other skills' SKILL.md files or the skills directory reveals prompt instructions, capabilities, and secrets that should be invisible to peer skills.

Content

Scanner excerpt · README.md (reported line 23)May include surrounding context.

bash
# Create skills directory if it doesn't exist
mkdir -p ~/.claude/skills/dhmz-weather

# Copy the skill file
cp SKILL.md ~/.claude/skills/dhmz-weather/

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
93% confidence
Finding

The skill explicitly instructs the agent to infer a city from conversation context, including the user's location, and to do so automatically without notice. This creates a privacy-sensitive behavior where location-derived data may be used or exposed without clear user consent, even though the downstream action is only a weather lookup.

Content

No source excerpt is available for this finding.

Autonomous Decision Making

Medium
Category
Excessive Agency
Confidence
91% confidence
Finding

The skill directs the agent to act autonomously and 'not ask the user,' including using inferred context to choose a city and immediately making network requests. While low-stakes in a weather skill, this reduces user control and can compound privacy issues by turning ambiguous context into action without confirmation.

Content

Scanner excerpt · SKILL.md (reported line 18)May include surrounding context.

md
1. **If a city is provided as argument** (e.g., `/dhmz-weather Zagreb`): Immediately fetch and display weather for that city
2. **If no city is provided**: Infer the city from conversation context (user's location, previously mentioned cities, or project context). If no context available, default to **Zagreb** (capital city)

**Do not ask the user what they want** - just fetch the weather data immediately and present it in a readable format.

## Weather Emojis

Natural-Language Policy Violations

Low
Category
Not specified by scanner
Confidence
84% confidence
Finding

The file states that all data is in Croatian, which creates a language constraint in the skill behavior. There is no accompanying note that the skill can translate, summarize in the user's preferred language, or otherwise offer a language choice.

Content

No source excerpt is available for this finding.

Static analysis

No suspicious patterns detected.